A smart attendance system based on regional terminals

Abstract

The present invention provides a smart attendance system based on regional terminals, which belongs to the field of access control management technology, and includes a user data collection unit, an administrator authorized user and data binding unit, a data management unit, a visiting user application entry and exit and information collection unit, a visiting user approval and virtual card generation unit, an attendance machine access data storage unit, a data sending cache and distribution management unit, an access control unit, a card swiping machine unit, and a face recognition and attendance point unit. The present invention can automatically perform face recognition attendance in the attendance area of ​​the company, avoiding the occurrence of contact and queuing problems such as some employees forgetting to punch in and fingerprint punching, and can perform face recognition on the position personnel at the same time, and can identify whether the visiting personnel is consistent with the original applicant, which has higher security and convenience, and avoids the problem of access cards being stolen or borrowed.

Description

Technical Field

[0001] The present invention relates to the technical field of access control management, and in particular to a smart attendance system based on regional terminals. Background Art

[0002] At present, the access to technical security access control system equipment is generally through SDK access or network protocol access. However, due to the lack of a unified access method, different brands or even different models of access control equipment of the same brand use different SDKs or protocols.

[0003] At present, the access control management system loads the SDK or communication protocol corresponding to the access control device controller, and connects the access control device point-to-point. However, since the general access control devices are deployed in each independent area, each access control device is generally located in an independent network segment; for security reasons, there is usually a firewall between the access control management system and each access control controller. If the cross communication method is adopted, that is, the one-to-many relationship between users and devices, it will cause confusion in network management. At the same time, the access control device generally has an upper limit for communication connections. If the number of accounts connected to the same access control device reaches the upper limit, other pre-logged accounts need to wait for the account that has logged in to the access control management system to log out before they can log in. When the management system client, device model, and number of devices increase significantly, the communication and operation between the access control management system and the access control devices will be very complicated.

[0004] The existing access control systems basically use a single access card or a single QR code recognition mode. Since the IC can be borrowed or stolen at any time, the identity of the person entering cannot be uniquely determined. At the same time, during rush hour, there are often queues for clocking in, which is very time-consuming and affects the efficiency of staff passing through the access control. Therefore, it is necessary to design a safer and more automated integrated system that combines access control and attendance. Summary of the invention

[0005] The purpose of the present invention is to provide a smart attendance system based on regional terminals to solve the technical problems mentioned in the background technology.

[0006] In order to achieve the above object, the technical solution adopted by the present invention is as follows:

[0007] A smart attendance system based on regional terminals, including a user data collection unit, an administrator authorized user and data binding unit, a data management unit, an access user application entry and exit and information collection unit, an access user approval and virtual card generation unit, an attendance machine access data storage unit, a data sending cache and distribution management unit, an access control unit, a card swiping machine unit and a face recognition and attendance point unit. The user data collection unit is connected to the administrator authorized user and data binding unit. The user data collection unit is used to collect the face information, basic information and entry and exit cards of the unit staff, and the face information, basic information and entry and exit cards of the staff are stored in the card reader. The card information is sent to the administrator's authorized user and data binding unit for approval. The administrator's authorized user and data binding unit are connected to the data management unit. The administrator's authorized user and data binding unit are used to approve the employee information passed in by the user data collection unit and assign permissions. The approved information and permission information are sent to the data management unit for storage. The visiting user application for entry and exit and information collection unit is connected to the visiting user approval and virtual card generation unit. The visiting user application for entry and exit and information collection unit is used for external visitors to apply for a visit, submit visit information materials and collect the visiting personnel's facial information, and pass it to the visiting user approval and virtual card generation unit. The access user approval and virtual card generation unit approves the access information transmitted by the access user application and information collection unit, and returns the access code to the access user application and information collection unit after the approval. The access user approval and virtual card generation unit is connected to the access data storage unit of the attendance machine. The access user approval and virtual card generation unit transmits the approved visitor's information and access code to the data management unit. The data management unit is connected to the access data storage unit of the attendance machine. The data management unit stores the facial information of the staff and the visitor, the basic information of the staff and the visitor, the staff access card information and the access code The data is transmitted to the attendance machine access data storage unit, the attendance machine access data storage unit is connected to the external OA system unit, the attendance machine access data storage unit is connected to the data sending cache and distribution management unit, the attendance machine access data storage unit transmits the facial information of the employee and the visitor, the employee access card information and the access code to the data sending cache and distribution management unit, the data sending cache and distribution management unit are respectively connected to the access control unit, the card swiping machine unit and the face recognition and attendance point unit, the data sending cache and distribution management unit controls the access control unit to open according to the identification information sent back by the card swiping machine unit and the face recognition and attendance point unit.

[0008] Furthermore, the user data collection unit includes an ID card recognition module, an access card generation module and a face image collection module. The ID card recognition module is used to recognize the ID card information of the employee and connect with the external information system interface to authenticate the ID card information. The access card generation module includes an access card issuing machine, an access card writing device and a touch screen. The touch screen is used to input the employee's authority information, and then the access card writing device writes the authority information into the access card, and the access card issuing machine spits out the access card. The face image collection module is used to collect the employee's face information.

[0009] Furthermore, the visiting user application entry and exit and information collection unit includes a visiting user application interface module, an ID card information recognition module, a virtual card acquisition module and a visiting user face collection module. The visiting user application interface module is used for the visiting user to input access information and ID card information, and transmit the ID card information to the ID card information recognition module. The ID card information recognition module is connected to the external information system interface and passes the ID card information to be authenticated. The visiting user face collection module is used to collect the face information of the visiting user, and the virtual card acquisition module is used to obtain the access code returned after approval.

[0010] Furthermore, the face recognition and attendance point unit includes a face recognition camera station, a face recognition camera cluster construction module, a face recognition range calculation module and a camera station face recognition number allocation module. The face recognition range calculation module and the camera station face recognition number allocation module are used to design a number of face recognition camera stations according to the face attendance location and the access control location. The face recognition camera cluster construction module is connected to the face recognition range calculation module. The face recognition camera cluster construction module constructs a face recognition camera cluster according to the location information and recognition radius of the face recognition camera station. The face recognition range calculation module is connected to the camera station face recognition number allocation module. The face recognition range calculation module is used to calculate the face recognition range for constructing a face recognition camera cluster. The face recognition camera cluster construction module is used to allocate the number of people to be recognized according to the recognition capability of each face recognition camera station. The face recognition camera station includes a number of cameras, and one attendance point is provided with more than two cameras.

[0011] Furthermore, the process of face recognition and face recognition by the attendance point unit includes the following steps:

[0012] Step 1: Design several face recognition camera stations. The number of face recognition camera stations is proportional to the width of the location of the face attendance and the access control location. The camera surface of each face recognition camera station is on the same plane. The camera surfaces of the face recognition camera stations are parallel or at an angle of less than 45°. Construct a face recognition camera cluster based on the location information and recognition radius of the face recognition camera station.

[0013] Step 2: Calculate the face recognition range of the face recognition camera cluster;

[0014] Step 3: Allocate the number of people to be recognized based on the recognition capability of each facial recognition camera station;

[0015] Step 4: Identify the assigned number of people and upload the identification results to the data delivery cache and distribution management unit.

[0016] Furthermore, the specific process of constructing the face recognition camera cluster in step 1 is as follows:

[0017] Step 1.1: Given a face recognition camera station object dataset of size n, let I = 1 and select k initial cluster centers Z j (I), j = 1, 2, 3, ..., k, k is a positive integer;

[0018] Step 1.2: Calculate the distance D(x) between each face recognition camera station and the cluster center i ,Z j (I)), i = 1, 2, 3, ..., k, if D(x i ,Z k (I)) = min{D(x i ,Z j (I)),i=1,2,3,…n} then x i ∈C k ;

[0019] Step 1.3: Calculate k new cluster centers, j = 1, 2, 3..., k, that is, the distance of each dimension of all elements in the cluster;

[0020] Step 1.4: Determine Z j (I+1)≠Z j (I), j = 1, 2, 3, ..., k, then I = I + 1, return to step 2.2, otherwise proceed to the next step;

[0021] Step 1.5: Determine the clustering distance of the face recognition camera station. The specific formula is as follows:

[0022]

[0023] Among them, x1 is the X coordinate constructed by the recognition point, x2 is the Y coordinate constructed by the recognition point, and x3 is the average radiation radius. Substituting x1, x2 and x3 into the K-means algorithm for calculation, the clustering result obtained is the face recognition camera cluster of the face recognition camera station.

[0024] Furthermore, the specific process of step 2 is:

[0025] Step 2.1: Calculate the straight-line distance between two face recognition camera stations, where the distance is the distance of the recognition surface;

[0026] Step 2.2: Calculate the breakpoint of the face recognition camera station; the formula is:

[0027]

[0028] Among them, L ik is the distance from the i-th face recognition camera station to the breakpoint k; M i 、M j For the recognition performance of face recognition camera stations i and j, the average daily face recognition volume of the face recognition camera station is set to be expressed, and the initial setting attendance volume is used to express it. ij is the distance between face recognition camera stations i and j;

[0029] Step 2.3: Calculate the recognition surface field strength of the face recognition camera station. The formula is:

[0030]

[0031] Among them, F ik is the face recognition speed of the i-th face recognition camera station at point k;

[0032] Step 2.4: Calculate the recognition radius of the face recognition camera station. The formula is:

[0033]

[0034] Among them, R ik is the radiation radius of the i-th face recognition camera station at point k;

[0035] Step 2.5: Calculate the average recognition radius of the face recognition camera station. The formula is:

[0036]

[0037] Further, in step 3, the allocation indicator system for allocating the number of people to be recognized includes the face recognition capability of the face recognition camera station, the utilization rate of the face recognition camera station and the status of the face recognition camera station. The face recognition capability of the face recognition camera station includes the average recognition radius and the average number of faces recognized per day. The utilization rate of the face recognition camera station includes the utilization rate of the camera image recognition running time and the ratio of the number of faces recognized simultaneously by the face recognition camera station to the number of top faces recognized simultaneously by the face recognition camera station. The status of the face recognition camera station includes the relative degree centrality, the relative proximity centrality and the relative eigenvector centrality.

[0038] The specific construction process of the camera image recognition running time utilization is as follows:

[0039]

[0040] Where P t represents the average operating time utilization of cameras at different face recognition camera stations, T m represents the cumulative running time of the face recognition camera station every day, M i Indicates the number of corresponding cameras, T d represents the planned operating time of each camera every day, and n represents the number of attendance checks per day;

[0041] The construction process of the ratio of the number of faces recognized simultaneously by the face recognition camera station to the number of top faces recognized simultaneously by the face recognition camera station is,

[0042]

[0043] Where Q t represents the average operating time utilization of cameras at different face recognition camera stations, Q m It indicates the average number of faces recognized by the face recognition camera station every day. is the ratio coefficient of the number of faces recognized by different types of cameras, T j Indicates the rated number of recognized faces of different cameras, M i Indicates the number of corresponding cameras, and N indicates the frequency of the corresponding camera processing operation;

[0044] The face recognition camera station obtains the relative degree centrality, relative proximity centrality and relative eigenvector centrality based on the number of times people pass by the position of the recognized face, and constructs a social network based on the frequency of people passing by the recognized face. First, a second-order occurrence network of face recognition camera stations and the number of times people pass by the recognized face is established. The second-order occurrence network forms an original network adjacency matrix R′, the number of face recognition camera stations is n, and the number of people appearing on the recognized face is m, where the matrix elements are,

[0045]

[0046] Where i = 1, 2, ..., n; j = 1, 2, ..., m;

[0047] The second-order occurrence network is converted into a first-order network for analysis. According to the relationship between the point pairs, the relationship between the face recognition camera station and the age of the person appearing on the recognition surface is converted into a first-order network matrix between the face recognition camera station and the face recognition camera station using the corresponding product method. The relationship between the face recognition camera station and the face recognition camera station measures the people of the same age group appearing on the recognition surface between the face recognition camera stations. Let the adjacency matrix of the network between the face recognition camera station and the face recognition camera station generated by the conversion to generate a first-order network matrix based on the competitive relationship of the age group structure be R h , where the elements are Defined as the age structure competition coefficient of the face recognition camera station personnel, the matrix R h The corresponding network is called the face recognition camera station personnel age group structure competition relationship subnetwork;

[0048] Point degree centrality is the number of points in the network that are directly connected to the point. It is used to measure the central position of an individual in the network, that is, the number of people who appear on the recognition surface of the face recognition camera station. It is divided into absolute degree centrality and relative degree centrality. The relative degree centrality is convenient for comparison between networks of different sizes. Its formula is expressed as:

[0049]

[0050] In the formula, g is the network consisting of g nodes, n i represents the i-th node, n j represents the jth node;

[0051] Closeness centrality measures the centrality of a node by calculating the distance between the node and other nodes. The closeness centrality of a node in a network is inversely proportional to the sum of the shortcut distances to other nodes. If a point is connected to other points through a short path, then the point has good closeness centrality. The relative closeness centrality formula is expressed as:

[0052]

[0053] In the formula, g is the network composed of g nodes, n i represents the i-th node, n j represents the jth node;

[0054] The eigenvector centrality measures the value of a point by the importance of its neighboring points, emphasizing the value of a node in the network. The value of a node in the network is determined by its neighboring points. The relative eigenvector formula is expressed as:

[0055]

[0056] In the formula, g is the network composed of g points, r ij is the adjacency matrix of the network, and λ is a constant;

[0057] Establish a second-order data network of face recognition camera stations and face data stored in face recognition camera stations, and form an original network adjacency matrix R″ from the second-order data network. The number of face recognition camera stations is n, and the number of face data is m. Transform and generate a first-order network of face recognition camera stations and face recognition camera station data based on face data competition. The adjacency matrix is ​​R k , set the face recognition camera station face data competition coefficient, the matrix R kThe corresponding network is called the face recognition camera station face data resource competition relationship subnetwork. The relative point degree centrality of the face recognition camera station in the face recognition camera station face data resource competition relationship subnetwork is also calculated according to the formula Solve it, and the relative eigenvector centrality is calculated according to the formula To solve.

[0058] Furthermore, the characteristic function of the allocation game for allocating the number of identified people is:

[0059]

[0060] Among them, x s represents the total number of joint hybrid strategies of all cameras, x N / S Indicates the totality of all cameras in N / S or strategies, u i (x,y) represents the expected distribution of face ratios for player i corresponding to strategy combination x,y, that is, the proportion of people appearing at the attendance point at the same time. For the sub-alliance S∈P(N) in the n-camera alliance game, the maximum total alliance utility that can be obtained after the alliance members cooperate is recorded as v(S). All v(S) that an alliance can give constitute a real-valued function of P(N)→R;

[0061] The characteristic function of the allocation is expressed as:

[0062]

[0063] in, In It is the competition coefficient between face recognition camera stations based on the face data resources of face recognition camera stations, that is, the face recognition camera stations in the sub-alliance are connected to each other through competition for face data resources, which determines the degree of utility contribution of the face recognition camera station to the face recognition camera station sub-alliance. represents the resource utility of face recognition camera station i and, R ij is the standardized value of the utility index, R i1 Represents the recognition radius of the face recognition camera station itself, R i2 Represents the number of facial recognition cameras recognized per day. represents the recognition level of face recognition camera station i, S ij is the standardized value of the utility indicator, including the utilization rate of the camera image recognition running time of the face recognition camera station, the ratio of the number of faces recognized simultaneously by the face recognition camera station to the number of top faces recognized simultaneously by the face recognition camera station, represents the network structure utility of face recognition camera station i, C ij is the standardized value of utility, C i1Representative face recognition camera station social network structure utility recognition face personnel age group competition subnetwork relative degree centrality, C i2 Represents the relative closeness centrality in the competitive subnetwork of the age group of the identified face personnel, C i3 represents the relative eigenvector centrality in the competitive subnetwork of the age group of the identified face, C i4 represents the relative degree centrality in the face data resource competition subnetwork, C i5 Represents the relative feature vector centrality in the face data resource competition subnetwork. It should be noted that when i is an isolated node in the sub-alliance network, the alliance utility is 0.

[0064] Furthermore, the process of assigning the number of people to be identified is the process of assigning Shapely values, and the Shapely value gives the cooperative game<N,v> In the formula, the utility of camera i of player i is the expected value of the marginal contribution of camera i to each camera sub-alliance it participates in. The Shapely value of player i is calculated as follows:

[0065]

[0066] The face recognition camera stations of the game parties form sub-alliances in a random order, and the probability of any order occurring is the same, which is Camera i of the game player forms an alliance S with the (|S|-1) cameras in front of it. Then the marginal contribution of camera i to the alliance is v(S)-v(Si). There are (|S|-1)! (n-|S|)! ways to arrange the cameras of the game player of S / i and N / S successively. Therefore, the probability of forming each alliance is The final product is the value of the number of people assigned to i, which is the percentage value at a time point;

[0067] The process of allocating the number of people to be recognized is to input the collaborative recognition alliance matrix and the original data of the alliance utility indicators, standardize the original data of the utility indicators, extract members from the collaborative recognition alliance to form all cooperative game face recognition camera station combinations, calculate the utility of each combination through the characteristic function, and store it in the combination, calculate the cumulative sum of the Shapely values ​​of each face recognition camera station in each combination, and calculate the Shapely value of each face recognition camera station through the Shapely value benefit distribution function.

[0068] This system will replace the original various cards and certificates of the enterprise, meet the work, life and daily management of employees, be more humane in service, and reduce the continuous investment of the enterprise. The cost of the original enterprise card certificate has the problem of repeated investment, and this system reduces the management cost. The bus integration technology and information integration technology are used to integrate the permissions of multiple management subsystems such as the smart park, and the information resource sharing of multiple systems is realized through the internal network, so as to improve the management convenience for the enterprise and provide more considerate services for employees.

[0069] To open the access control, the face and the recognized access card must be the same before the access control is opened, avoiding the problem of the access card being stolen or borrowed.

[0070] The present invention has the following beneficial effects due to the adoption of the above technical solution:

[0071] The present invention can automatically perform face recognition attendance for people entering the company's attendance area, thereby avoiding the occurrence of contact and queuing problems such as some employees forgetting to punch in and fingerprint punching. At the same time, face recognition can be performed on the opposite personnel, and it can be identified whether the visiting personnel is consistent with the original applicant. The invention has higher security and convenience, and avoids the problem of access cards being stolen or borrowed. At the same time, through a high-precision allocation algorithm, each camera can be brought into play in the optimal state, and reasonable allocation of attendance points with a large number of people is realized, and attendance is fast and accurate. Attendance can be performed for people with different heights or side faces entering and exiting the attendance points, and the attendance effect solves the defects of the existing face attendance. BRIEF DESCRIPTION OF THE DRAWINGS

[0072] Figure 1 is a system block diagram of the present invention;

[0073] Figure 2 It is a module block diagram of the user data acquisition unit of the present invention;

[0074] Figure 3 It is the access user application entry and exit and information collection unit of the present invention;

[0075] Figure 4 It is a schematic diagram of the camera station structure of the present invention;

[0076] Figure 5 It is a block diagram of the face recognition and attendance point unit module of the present invention;

[0077] Figure 6 It is a flow chart of the automatic face recognition process of the present invention. DETAILED DESCRIPTION

[0078] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below with reference to the accompanying drawings and preferred embodiments. However, it should be noted that many details listed in the specification are only for the purpose of enabling the reader to have a thorough understanding of one or more aspects of the present invention, and these aspects of the present invention can be implemented even without these specific details.

[0079] like Figure 1 As shown, a smart attendance system based on regional terminals includes a user data collection unit, an administrator authorized user and data binding unit, a data management unit, an access user application entry and exit and information collection unit, an access user approval and virtual card generation unit, an attendance machine access data storage unit, a data sending cache and distribution management unit, an access control unit, a card swiping machine unit and a face recognition and attendance point unit. The user data collection unit is connected to the administrator authorized user and data binding unit. The user data collection unit is used to collect the face information, basic information and distribution entry and exit cards of the unit employees, and the face information, basic information and The access card information is sent to the administrator's authorized user and data binding unit for approval. The administrator's authorized user and data binding unit are connected to the data management unit. The administrator's authorized user and data binding unit are used to approve the employee information passed in by the user data collection unit and assign permissions. The approved information and permission information are sent to the data management unit for storage. The visiting user application for access and information collection unit is connected to the visiting user approval and virtual card generation unit. The visiting user application for access and information collection unit is used for external visitors to apply for access, submit access information materials and collect the visiting personnel's facial information, and pass it to the visiting user approval and virtual card generation unit. The access user approval and virtual card generation unit approves the access information transmitted by the access user application and information collection unit, and returns the access code to the access user application and information collection unit after the approval. The access user approval and virtual card generation unit is connected to the access data storage unit of the attendance machine. The access user approval and virtual card generation unit transmits the approved visitor's information and access code to the data management unit. The data management unit is connected to the access data storage unit of the attendance machine. The data management unit stores the facial information of the staff and the visitor, the basic information of the staff and the visitor, the staff access card information and the access code of the staff. The code is transmitted to the attendance machine access data storage unit, the attendance machine access data storage unit is connected to the external OA system unit, the attendance machine access data storage unit is connected to the data sending cache and distribution management unit, the attendance machine access data storage unit transmits the facial information of the employee and the visitor, the employee access card information and the access code to the data sending cache and distribution management unit, the data sending cache and distribution management unit are respectively connected to the access control unit, the card swiping machine unit and the face recognition and attendance point unit, the data sending cache and distribution management unit controls the access control unit to open according to the identification information sent back by the card swiping machine unit and the face recognition and attendance point unit.

[0080] The user data collection unit is used to enter the information of new employees in the company, or the information of all employees when the system is just installed. The system can be used in areas such as tobacco companies, and can more safely and automatically control the entire park area of ​​the company.

[0081] The administrator authorizes users and data binding units, which are generally approved by the personnel management and general manager. After the personnel approval, the general manager will approve it. According to different permissions, you can enter different areas, that is, inside the formula, such as the elevator, and some floors cannot be accessed without permission. There are also camera stations in the stairs for automatic face recognition, which means that when people get on the elevator, they need to swipe the access card, and the person who swipes the card needs to match the face before they can get on.

[0082] The data management unit is a server that cannot be located from the outside and can only transmit data outward, which is more secure. The access user application entry and information collection unit is generally installed on the mobile phone web page or WeChat applet mode for user application. The access user approval and virtual card generation unit is for human resources leaders to approve. When approving, it is necessary to check with the access department of the internal OA system before approval. The attendance machine access data storage unit is mainly used as an intermediate data storage server and can be located. The data delivery cache and distribution management unit is managed as a controller to manage the distribution and control of memory data.

[0083] In the embodiment of the present invention, Figure 2 As shown, the user data acquisition unit includes an ID card recognition module, an access card generation module and a face image acquisition module. The ID card recognition module is used to recognize the ID card information of the employee and connect with the external information system interface to authenticate the ID card information. The access card generation module includes an access card issuing machine, an access card writing device and a touch screen. The touch screen is used to input the employee's authority information, and then the access card writing device writes the authority information into the access card, and the access card issuing machine spits out the access card. The face image acquisition module is used to collect the employee's face information.

[0084] In the embodiment of the present invention, Figure 3 As shown, the visiting user application entry and exit and information collection unit includes a visiting user application interface module, an ID card information recognition module, a virtual card acquisition module and a visiting user face collection module. The visiting user application interface module is used for the visiting user to input access information and ID card information, and transmit the ID card information to the ID card information recognition module. The ID card information recognition module is connected to the external information system interface and passes the ID card information to be authenticated. The visiting user face collection module is used to collect the face information of the visiting user, and the virtual card acquisition module is used to obtain the access code returned after approval.

[0085] In the embodiment of the present invention, Figure 4-5 As shown, the face recognition and attendance point unit includes a face recognition camera station, a face recognition camera cluster construction module, a face recognition range calculation module and a camera station face recognition number allocation module. The face recognition range calculation module and the camera station face recognition number allocation module are used to design a number of face recognition camera stations according to the location of face attendance and the access control location. The face recognition camera cluster construction module is connected to the face recognition range calculation module. The face recognition camera cluster construction module constructs a face recognition camera cluster according to the location information and recognition radius of the face recognition camera station. The face recognition range calculation module is connected to the camera station face recognition number allocation module. The face recognition range calculation module is used to calculate the face recognition range for constructing the face recognition camera cluster. The face recognition camera cluster construction module is used to allocate the number of recognized people according to the recognition capability of each face recognition camera station. The face recognition camera station includes a number of cameras, and one attendance point is provided with more than two cameras.

[0086] The process of face recognition and attendance point unit face recognition includes the following steps:

[0087] Step 1: Design several facial recognition camera stations. The number of facial recognition camera stations is proportional to the width of the facial attendance location and the access control location. The camera surface of each facial recognition camera station is on the same plane. The camera surfaces of the facial recognition camera stations are parallel or have an angle of less than 45°. Construct a facial recognition camera cluster based on the location information and recognition radius of the facial recognition camera station.

[0088] The specific process of building a face recognition camera cluster in step 1 is as follows:

[0089] Step 1.1: Given a face recognition camera station object dataset of size n, let I = 1 and select k initial cluster centers Z j (I), j = 1, 2, 3, ..., k, k is a positive integer;

[0090] Step 1.2: Calculate the distance D(x) between each face recognition camera station and the cluster center i ,Z j (I)), i = 1, 2, 3, ..., k, if D(x i ,Z k (I)) = min{D(x i ,Z j (I)),i=1,2,3,…n} then x i ∈C k ;

[0091] Step 1.3: Calculate k new cluster centers, j = 1, 2, 3..., k, that is, the distance of each dimension of all elements in the cluster;

[0092] Step 1.4: Determine Z j (I+1)≠Z j (I), j = 1, 2, 3, ..., k, then I = I + 1, return to step 2.2, otherwise proceed to the next step;

[0093] Step 1.5: Determine the clustering distance of the face recognition camera station. The specific formula is as follows:

[0094]

[0095] Among them, x1 is the X coordinate constructed by the recognition point, x2 is the Y coordinate constructed by the recognition point, and x3 is the average radiation radius. Substituting x1, x2 and x3 into the K-means algorithm for calculation, the clustering result obtained is the face recognition camera cluster of the face recognition camera station.

[0096] Step 2: Calculate the face recognition range for building the face recognition camera cluster.

[0097] Step 2.1: Calculate the straight-line distance between two face recognition camera stations, where the distance is the distance of the recognition surface;

[0098] Step 2.2: Calculate the breakpoint of the face recognition camera station; the formula is:

[0099]

[0100] Among them, L ik is the distance from the i-th face recognition camera station to the breakpoint k; M i 、M j For the recognition performance of face recognition camera stations i and j, the average daily face recognition volume of the face recognition camera station is set to be expressed, and the initial setting attendance volume is used to express it. ij is the distance between face recognition camera stations i and j;

[0101] Step 2.3: Calculate the recognition surface field strength of the face recognition camera station. The formula is:

[0102]

[0103] Among them, F ik is the face recognition speed of the i-th face recognition camera station at point k;

[0104] Step 2.4: Calculate the recognition radius of the face recognition camera station. The formula is:

[0105]

[0106] Among them, Rik is the radiation radius of the i-th face recognition camera station at point k;

[0107] Step 2.5: Calculate the average recognition radius of the face recognition camera station. The formula is:

[0108]

[0109] Step 3: Assign the number of people to be recognized based on the recognition capability of each facial recognition camera station.

[0110] In step 3, the allocation indicator system for allocating the number of people to be recognized includes the face recognition capability of the face recognition camera station, the utilization rate of the face recognition camera station and the status of the face recognition camera station. The face recognition capability of the face recognition camera station includes the average recognition radius and the average number of faces recognized per day. The utilization rate of the face recognition camera station includes the utilization rate of the camera image recognition running time and the ratio of the number of faces recognized simultaneously by the face recognition camera station to the number of top faces recognized simultaneously by the face recognition camera station. The status of the face recognition camera station includes the relative degree centrality, the relative proximity centrality and the relative eigenvector centrality.

[0111] The specific construction process of the camera image recognition running time utilization is as follows:

[0112]

[0113] Where P t represents the average operating time utilization of cameras at different face recognition camera stations, T m represents the cumulative running time of the face recognition camera station every day, M i Indicates the number of corresponding cameras, T d represents the planned operating time of each camera every day, and n represents the number of attendance checks per day;

[0114] The construction process of the ratio of the number of faces recognized simultaneously by the face recognition camera station to the number of top faces recognized simultaneously by the face recognition camera station is,

[0115]

[0116] Where Q t represents the average operating time utilization of cameras at different face recognition camera stations, Q m It indicates the average number of faces recognized by the face recognition camera station every day. is the ratio coefficient of the number of faces recognized by different types of cameras, T j Indicates the rated number of recognized faces of different cameras, M i Indicates the number of corresponding cameras, and N indicates the frequency of the corresponding camera processing operation;

[0117] The face recognition camera station obtains the relative degree centrality, relative proximity centrality and relative eigenvector centrality based on the number of times people pass by the position of the recognized face, and constructs a social network based on the frequency of people passing by the recognized face. First, a second-order occurrence network of face recognition camera stations and the number of times people pass by the recognized face is established. The second-order occurrence network forms an original network adjacency matrix R′, the number of face recognition camera stations is n, and the number of people appearing on the recognized face is m, where the matrix elements are,

[0118]

[0119] Where i = 1, 2, ..., n; j = 1, 2, ..., m;

[0120] The second-order occurrence network is converted into a first-order network for analysis. According to the relationship between the point pairs, the relationship between the face recognition camera station and the age of the person appearing on the recognition surface is converted into a first-order network matrix between the face recognition camera station and the face recognition camera station using the corresponding product method. The relationship between the face recognition camera station and the face recognition camera station measures the people of the same age group appearing on the recognition surface between the face recognition camera stations. Let the adjacency matrix of the network between the face recognition camera station and the face recognition camera station generated by the conversion to generate a first-order network matrix based on the competitive relationship of the age group structure be R h , where the elements are Defined as the age structure competition coefficient of the face recognition camera station personnel, the matrix R h The corresponding network is called the face recognition camera station personnel age group structure competition relationship subnetwork;

[0121] Point degree centrality is the number of points in the network that are directly connected to the point. It is used to measure the central position of an individual in the network, that is, the number of people who appear on the recognition surface of the face recognition camera station. It is divided into absolute degree centrality and relative degree centrality. The relative degree centrality is convenient for comparison between networks of different sizes. Its formula is expressed as:

[0122]

[0123] In the formula, g is the network consisting of g nodes, n i represents the i-th node, n j represents the jth node;

[0124] Closeness centrality measures the centrality of a node by calculating the distance between the node and other nodes. The closeness centrality of a node in a network is inversely proportional to the sum of the shortcut distances to other nodes. If a point is connected to other points through a short path, then the point has good closeness centrality. The relative closeness centrality formula is expressed as:

[0125]

[0126] In the formula, g is the network composed of g nodes, n i represents the i-th node, nj represents the j-th node;

[0127] The eigenvector centrality measures the value of a point by the importance of its neighboring points, emphasizing the value of a node in the network. The value of a node in the network is determined by its neighboring points. The relative eigenvector formula is expressed as:

[0128]

[0129] In the formula, g is the network composed of g points, r ij is the adjacency matrix of the network, and λ is a constant;

[0130] Establish a second-order data network of face recognition camera stations and face data stored in face recognition camera stations, and form an original network adjacency matrix R″ from the second-order data network. The number of face recognition camera stations is n, and the number of face data is m. Transform and generate a first-order network of face recognition camera stations and face recognition camera station data based on face data competition. The adjacency matrix is ​​R k , set the face recognition camera station face data competition coefficient, the matrix R k The corresponding network is called the face recognition camera station face data resource competition relationship subnetwork. The relative point degree centrality of the face recognition camera station in the face recognition camera station face data resource competition relationship subnetwork is also calculated according to the formula Solve it, and the relative eigenvector centrality is calculated according to the formula To solve.

[0131] The characteristic function of the allocation game for allocating the number of identified people is:

[0132]

[0133] Among them, x s represents the total number of joint hybrid strategies of all cameras, x N / S Indicates the totality of all cameras in N / S or strategies, u i (x,y) represents the expected distribution of face ratios for player i corresponding to strategy combination x,y, that is, the proportion of people appearing at the attendance point at the same time. For the sub-alliance S∈P(N) in the n-camera alliance game, the maximum total alliance utility that can be obtained after the alliance members cooperate is recorded as v(S). All v(S) that an alliance can give constitute a real-valued function of P(N)→R;

[0134] The characteristic function of the allocation is expressed as:

[0135]

[0136] in, In It is the competition coefficient between face recognition camera stations based on the face data resources of face recognition camera stations, that is, the face recognition camera stations in the sub-alliance are connected to each other through competition for face data resources, which determines the degree of utility contribution of the face recognition camera station to the face recognition camera station sub-alliance. represents the resource utility of face recognition camera station i and, R ij is the standardized value of the utility index, R i1 Represents the recognition radius of the face recognition camera station itself, R i2 Represents the number of facial recognition cameras recognized per day. represents the recognition level of face recognition camera station i, S ij is the standardized value of the utility indicator, including the utilization rate of the camera image recognition running time of the face recognition camera station, the ratio of the number of faces recognized simultaneously by the face recognition camera station to the number of top faces recognized simultaneously by the face recognition camera station, represents the network structure utility of face recognition camera station i, C ij is the standardized value of utility, C i1 Represents the social network structure utility of face recognition camera station, recognition face, age group, relative degree centrality in the competitive subnetwork, C i2 Represents the relative closeness centrality in the competitive subnetwork of the age group of the identified face personnel, C i3 represents the relative eigenvector centrality in the competitive subnetwork of the age group of the identified face, C i4 represents the relative degree centrality in the face data resource competition subnetwork, C i5 Represents the relative feature vector centrality in the face data resource competition subnetwork. It should be noted that when i is an isolated node in the sub-alliance network, the alliance utility is 0.

[0137] The process of assigning the number of people to be identified is the process of assigning Shapely values. Shapely values ​​give the cooperative game<N,v> In the formula, the utility of camera i of player i is the expected value of the marginal contribution of camera i to each camera sub-alliance it participates in. The Shapely value of player i is calculated as follows:

[0138]

[0139] The face recognition camera stations of the game parties form sub-alliances in a random order, and the probability of any order occurring is the same, which is Camera i of the game player forms an alliance S with the (|S|-1) cameras in front of it. Then the marginal contribution of camera i to the alliance is v(S)-v(Si). There are (|S|-1)! (n-|S|)! ways to arrange the cameras of the game player of S / i and N / S successively. Therefore, the probability of forming each alliance is The final product is the value of the number of people assigned to i, which is the percentage value at a time point;

[0140] The process of allocating the number of people to be recognized is to input the collaborative recognition alliance matrix and the original data of the alliance utility indicators, standardize the original data of the utility indicators, extract members from the collaborative recognition alliance to form all cooperative game face recognition camera station combinations, calculate the utility of each combination through the characteristic function, and store it in the combination, calculate the cumulative sum of the Shapely values ​​of each face recognition camera station in each combination, and calculate the Shapely value of each face recognition camera station through the Shapely value benefit distribution function.

[0141] Step 4: Identify the number of people assigned to be identified, and upload the identification results to the data delivery cache and distribution management unit. The identified face data is judged to be an employee of the company. At the same time, if it is within the time range of attendance, it is also the first time to be identified, that is, the first time is set as attendance identification, and the others are not considered as attendance identification. Then if it is a visitor, after face recognition, the access control can be opened only after the face matches the swipe card, making access control management safer.

[0142] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A smart attendance system based on regional terminals, characterized by: It includes a user data collection unit, an administrator authorized user and a data binding unit, a data management unit, a visiting user application entry and exit and information collection unit, a visiting user approval and virtual card generation unit, an attendance machine access data storage unit, a data sending cache and distribution management unit, an access control unit, a card swiping machine unit and a face recognition and attendance point unit. The user data collection unit is connected to the administrator authorized user and the data binding unit. The user data collection unit is used to collect the face information, basic information and distribution entry and exit cards of the unit employees, and send the face information, basic information and entry and exit card information of the employees to the administrator authorized user and the data binding unit for approval. The administrator authorized user and the data binding unit are connected to the data management unit. The administrator authorized user and the data binding unit are used to approve the employee information transmitted by the user data collection unit and allocate authority, and send the approved information and authority information to the data management unit for storage. The visiting user application entry and exit and information collection unit is connected to the visiting user approval and virtual card generation unit. The visiting user application entry and exit and information collection unit is used for external visitors to apply for access, submit access information materials and collect the face information of the visiting personnel, and transmit it to the visiting user approval and virtual card generation unit. The approval and virtual card generation unit examines the access information transmitted by the access user application and information collection unit, and returns the access code to the access user application and information collection unit after the approval. The access user approval and virtual card generation unit is connected to the access data storage unit of the attendance machine. The access user approval and virtual card generation unit transmits the approved visitor's information and access code to the data management unit. The data management unit is connected to the access data storage unit of the attendance machine. The data management unit transmits the facial information of the staff and the visitor, the basic information of the staff and the visitor, the staff access card information and the access code to the attendance machine. The time attendance machine accesses the data storage unit, the time attendance machine accesses the data storage unit and is connected to the external OA system unit, the time attendance machine accesses the data storage unit and is connected to the data sending cache and distribution management unit, the time attendance machine accesses the data storage unit and transmits the facial information of the staff and the visitor, the staff access card information and the access code to the data sending cache and distribution management unit, the data sending cache and distribution management unit are respectively connected to the access control unit, the card swiping machine unit and the face recognition and attendance point unit, the data sending cache and distribution management unit controls the access control unit to be opened according to the identification information sent back by the card swiping machine unit and the face recognition and attendance point unit; The user data acquisition unit includes an ID card recognition module, an access card generation module and a face image acquisition module. The ID card recognition module is used to recognize the ID card information of the employee and connect with the external information system interface to authenticate the ID card information. The access card generation module includes an access card issuing machine, an access card writing device and a touch screen. The touch screen is used to input the authority information of the employee, and then the access card writing device writes the authority information into the access card, and the access card issuing machine spits out the access card. The face image acquisition module is used to collect the face information of the employee. The access user application entry and exit and information collection unit includes an access user application interface module, an ID card information recognition module, a virtual card acquisition module and an access user face collection module. The access user application interface module is used for the access user to input access information and ID card information, and transmit the ID card information to the ID card information recognition module. The ID card information recognition module is connected to the external information system interface and transmits the ID card information for authentication. The access user face collection module is used to collect the face information of the access user, and the virtual card acquisition module is used to obtain the access code returned after approval; The face recognition and attendance point unit includes a face recognition camera station, a face recognition camera cluster construction module, a face recognition range calculation module and a camera station face recognition number allocation module. The face recognition range calculation module and the camera station face recognition number allocation module are used to design a number of face recognition camera stations according to the location of face attendance and the access control location. The face recognition camera cluster construction module is connected to the face recognition range calculation module. The face recognition camera cluster construction module constructs a face recognition camera cluster according to the location information and recognition radius of the face recognition camera station. The face recognition range calculation module is connected to the camera station face recognition number allocation module. The face recognition range calculation module is used to calculate the face recognition range of the face recognition camera cluster. The face recognition camera cluster construction module is used to allocate the number of recognitions according to the recognition capability of each face recognition camera station. The face recognition camera station includes a number of cameras, and one attendance point is provided with two or more cameras. The process of face recognition and attendance point unit face recognition includes the following steps: Step 1: Design several face recognition camera stations. The number of face recognition camera stations is proportional to the width of the location of the face attendance and the access control location. The camera surface of each face recognition camera station is on the same plane. The camera surfaces of the face recognition camera stations are parallel or at an angle of less than 45°. Construct a face recognition camera cluster based on the location information and recognition radius of the face recognition camera station. Step 2: Calculate the face recognition range of the face recognition camera cluster; Step 3: Allocate the number of people to be recognized based on the recognition capability of each facial recognition camera station; Step 4: Identify the number of people assigned to be identified, and upload the identification results to the data delivery cache and distribution management unit; The specific process of building a face recognition camera cluster in step 1 is as follows: Step 1.1: Given a face recognition camera station object dataset of size n, let I = 1 and select k initial cluster centers Z j (I), j = 1, 2, 3, ..., k, k is a positive integer; Step 1.2: Calculate the distance D(x) between each face recognition camera station and the cluster center i ,Z j (I)), i = 1, 2, 3, ..., k, if D(x i ,Z k (I)) = min{D(x i ,Z j (I)),i=1,2,3,…n} then x i ∈C k ; Step 1.3: Calculate k new cluster centers, That is, take the distance of each dimension of all elements in the cluster; Step 1.4: Determine Z j (I+1)≠Z j (I), j = 1, 2, 3, ..., k, then I = I + 1, return to step 2.2, otherwise proceed to the next step; Step 1.5: Determine the clustering distance of the face recognition camera station. The specific formula is as follows: Among them, x1 is the X coordinate of the recognition point, x2 is the Y coordinate of the recognition point, and x3 is the average radiation radius. Substitute x1, x2, and x3 into the K-means algorithm for calculation, and the clustering result is the face recognition camera cluster of the face recognition camera station; The specific process of step 2 is: Step 2.1: Calculate the straight-line distance between two face recognition camera stations, where the distance is the distance of the recognition surface; Step 2.2: Calculate the breakpoint of the face recognition camera station; the formula is: Among them, L ik is the distance from the i-th face recognition camera station to the breakpoint k; M i 、M j For the recognition performance of face recognition camera stations i and j, the average daily face recognition volume of the face recognition camera station is set to be expressed, and the initial setting attendance volume is used to express it. ij is the distance between face recognition camera stations i and j; Step 2.3: Calculate the recognition surface field strength of the face recognition camera station. The formula is: Among them, F ik is the face recognition speed of the i-th face recognition camera station at point k; Step 2.4: Calculate the recognition radius of the face recognition camera station. The formula is: Among them, R ik is the radiation radius of the i-th face recognition camera station at point k; Step 2.5: Calculate the average recognition radius of the face recognition camera station. The formula is: In step 3, the allocation indicator system for allocating the number of people to be recognized includes the face recognition capability of the face recognition camera station, the utilization rate of the face recognition camera station and the status of the face recognition camera station. The face recognition capability of the face recognition camera station includes the average recognition radius and the average number of faces recognized per day. The utilization rate of the face recognition camera station includes the utilization rate of the camera image recognition running time and the ratio of the number of faces recognized simultaneously by the face recognition camera station to the number of top faces recognized simultaneously by the face recognition camera station. The status of the face recognition camera station includes the relative degree centrality, the relative proximity centrality and the relative eigenvector centrality. The specific construction process of the camera image recognition running time utilization is as follows: Where P t represents the average operating time utilization of cameras at different face recognition camera stations, T m represents the cumulative running time of the face recognition camera station every day, M i Indicates the number of corresponding cameras, T d represents the planned operating time of each camera every day, and n represents the number of attendance checks per day; The construction process of the ratio of the number of faces recognized simultaneously by the face recognition camera station to the number of top faces recognized simultaneously by the face recognition camera station is, Where Q t represents the average operating time utilization of cameras at different face recognition camera stations, Q m It indicates the average number of faces recognized by the face recognition camera station every day. is the ratio coefficient of the number of faces recognized by different types of cameras, T j Indicates the rated number of recognized faces of different cameras, M i Indicates the number of corresponding cameras, and N indicates the frequency of the corresponding camera processing operation; The face recognition camera station obtains the relative degree centrality, relative proximity centrality and relative eigenvector centrality based on the number of times people pass by the position of the recognized face, and constructs a social network based on the frequency of people passing by the recognized face. First, a second-order occurrence network of face recognition camera stations and the number of times people pass by the recognized face is established. The second-order occurrence network forms an original network adjacency matrix R′, the number of face recognition camera stations is n, and the number of people appearing on the recognized face is m, where the matrix elements are, Where i = 1, 2, ..., n; j = 1, 2, ..., m; The second-order occurrence network is converted into a first-order network for analysis. According to the relationship between the point pairs, the relationship between the face recognition camera station and the age of the person appearing on the recognition surface is converted into a first-order network matrix between the face recognition camera station and the face recognition camera station using the corresponding product method. The relationship between the face recognition camera station and the face recognition camera station measures the people of the same age group appearing on the recognition surface between the face recognition camera stations. Let the adjacency matrix of the network between the face recognition camera station and the face recognition camera station generated by the conversion to generate a first-order network matrix based on the competitive relationship of the age group structure be R h , where the elements are Defined as the age structure competition coefficient of the face recognition camera station personnel, the matrix R h The corresponding network is called the face recognition camera station personnel age group structure competition relationship subnetwork; Point degree centrality is the number of points in the network that are directly connected to the point. It is used to measure the central position of an individual in the network, that is, the number of people who appear on the recognition surface of the face recognition camera station. It is divided into absolute degree centrality and relative degree centrality. The relative degree centrality is convenient for comparison between networks of different sizes. Its formula is expressed as: In the formula, g is the network consisting of g nodes, n i represents the i-th node, n j represents the jth node; Closeness centrality measures the centrality of a node by calculating the distance between the node and other nodes. The closeness centrality of a node in a network is inversely proportional to the sum of the shortcut distances to other nodes. If a point is connected to other points through a short path, then the point has good closeness centrality. The relative closeness centrality formula is expressed as: In the formula, g is the network consisting of g nodes, n i represents the i-th node, n j represents the jth node; The eigenvector centrality measures the value of a point by the importance of its neighboring points, emphasizing the value of a node in the network. The value of a node in the network is determined by its neighboring points. The relative eigenvector formula is expressed as: In the formula, g is the network composed of g points, r ij is the adjacency matrix of the network, and λ is a constant; Establish a second-order data network of face recognition camera stations and face data stored in face recognition camera stations, and form an original network adjacency matrix R″ from the second-order data network. The number of face recognition camera stations is n, and the number of face data is m. Transform and generate a first-order network of face recognition camera stations and face recognition camera station data based on face data competition. The adjacency matrix is ​​R k , set the face recognition camera station face data competition coefficient, the matrix R k The corresponding network is called the face recognition camera station face data resource competition relationship subnetwork. The relative point degree centrality of the face recognition camera station in the face recognition camera station face data resource competition relationship subnetwork is also calculated according to the formula Solve it, and the relative eigenvector centrality is calculated according to the formula To solve; The characteristic function of the allocation game for allocating the number of identified people is: Among them, x s represents the total number of joint hybrid strategies of all cameras, x N / S Indicates the totality of all cameras in N / S or strategies, u i (x,y) represents the expected distribution of face ratios for player i corresponding to strategy combination x,y, that is, the proportion of people appearing at the attendance point at the same time. For the sub-alliance S∈P(N) in the n-camera alliance game, the maximum total alliance utility that can be obtained after the alliance members cooperate is recorded as v(S). All v(S) that an alliance can give constitute a real-valued function of P(N)→R; The characteristic function of the allocation is expressed as: in, In It is the competition coefficient between face recognition camera stations based on the face data resources of face recognition camera stations, that is, the face recognition camera stations in the sub-alliance are connected to each other through competition for face data resources, which determines the degree of utility contribution of the face recognition camera station to the face recognition camera station sub-alliance. represents the resource utility of face recognition camera station i and, R ij is the standardized value of the utility index, R i1 Represents the recognition radius of the face recognition camera station itself, R i2 Represents the number of facial recognition cameras recognized per day. represents the recognition level of face recognition camera station i, S ij is the standardized value of the utility indicator, including the utilization rate of the camera image recognition running time of the face recognition camera station, the ratio of the number of faces recognized simultaneously by the face recognition camera station to the number of top faces recognized simultaneously by the face recognition camera station, represents the network structure utility of face recognition camera station i, C ij is the standardized value of utility, C i1 Represents the social network structure utility of face recognition camera station, recognition face, age group, relative degree centrality in the competitive subnetwork, C i2 Represents the relative closeness centrality in the competitive subnetwork of the age group of the identified face personnel, C i3 represents the relative eigenvector centrality in the competitive subnetwork of the age group of the identified face, C i4 represents the relative degree centrality in the face data resource competition subnetwork, C i5 Represents the relative feature vector centrality in the face data resource competition subnetwork. It should be noted that when i is an isolated node in the sub-alliance network, the alliance utility is 0.

2. According to claim 1, a smart attendance system based on regional terminals is characterized in that: The process of assigning the number of people to be identified is the process of assigning Shapely values. Shapely values ​​give the cooperative game<N,v> In the formula, the utility of camera i of player i is the expected value of the marginal contribution of camera i to each camera sub-alliance it participates in. The Shapely value of player i is calculated as follows: The face recognition camera stations of the game parties form sub-alliances in a random order, and the probability of any order occurring is the same, which is Camera i of the game player forms an alliance S with the (|S|-1) cameras in front of it. Then the marginal contribution of camera i to the alliance is v(S)-v(Si). There are (|S|-1)! (n-|S|)! ways to arrange the cameras of the game player of S / i and N / S successively. Therefore, the probability of forming each alliance is The final product is the value of the number of people assigned to i, which is the percentage value at a time point; The process of allocating the number of people to be recognized is to input the collaborative recognition alliance matrix and the original data of the alliance utility indicators, standardize the original data of the utility indicators, extract members from the collaborative recognition alliance to form all cooperative game face recognition camera station combinations, calculate the utility of each combination through the characteristic function, and store it in the combination, calculate the cumulative sum of the Shapely values ​​of each face recognition camera station in each combination, and calculate the Shapely value of each face recognition camera station through the Shapely value benefit distribution function.

Images