Intelligent access control intelligent management system based on face recognition

Abstract

The invention relates to the technical field of computer vision, in particular to an intelligent access control intelligent management system based on face recognition. The method comprises the following steps: a user management unit manages and maintains personal information and access authority of authorized users based on an information management module; the face recognition unit captures face images of people going in and out through a portrait acquisition module on the basis of user information of the user management unit, introduces an aerial fog influence variable to carry out data processing, and carries out identity verification through a convolutional neural network algorithm which introduces an illumination variable; the access control unit is used for controlling access control through a control module on the basis of an identity verification result of the face recognition unit; the communication integration unit establishes network connection through a wireless network module on the basis of an IP network protocol, so that stable data transmission among the components in the access control system is ensured. According to the design, through strict permission setting, it is ensured that only authorized personnel can enter a specific area, and the safety of the system is enhanced.

Description

Technical Field

[0001] The present invention relates to the field of computer vision technology, and in particular to a smart access control intelligent management system based on face recognition. Background Art

[0002] The smart access control intelligent management system based on face recognition is a highly integrated security solution. It uses advanced computer vision and artificial intelligence technologies, especially deep learning algorithms, to capture facial images of people entering or leaving a specific area through cameras, and compares them with the pre-registered user database in real time to automatically identify individuals and control access rights. The system is not limited to simple door opening and closing operations, but can also provide a series of value-added services such as visitor management, personnel flow statistics, abnormal behavior monitoring, and emergency response. Through the Internet of Things (IoT) technology, smart access control can be seamlessly connected with other security equipment, such as surveillance cameras, alarm systems, and smart home devices to form a comprehensive security protection network. At the same time, with the powerful processing power and big data analysis tools of the cloud computing platform, the system can achieve remote management and intelligent decision support to ensure the security and efficient operation of the community or corporate environment. In addition, in order to protect user privacy and data security, the system usually adopts high-level encryption technology and strict access control policies to prevent unauthorized access and personal information leakage. Overall, the smart access control intelligent management system based on facial recognition represents an important innovation in the field of modern security. It greatly improves the convenience and safety of access control and provides a more intelligent and humane service experience for people's life and work.

[0003] In the existing smart access control intelligent management system based on face recognition, the camera may not be able to clearly capture the face image in strong light or dim environment, resulting in recognition failure; and wearing hats, glasses, masks and other items may block facial features and affect recognition accuracy. Although technology is constantly improving, under certain conditions (such as twins with high similarity, large changes in facial expressions, changes in appearance caused by aging, etc.), the system may still have misidentification (mistaking different people for the same person) or rejection (failure to correctly identify legitimate users), which will reduce user experience and cause security issues. Face recognition involves the collection, storage and use of personal biometric data. Once this sensitive information is leaked, it may pose a serious threat to user privacy. Therefore, a smart access control intelligent management system based on face recognition is designed. Summary of the invention

[0004] The purpose of the present invention is to provide a smart access control intelligent management system based on face recognition, so as to solve the problem raised in the above background technology that in the existing smart access control intelligent management system based on face recognition, the camera may not be able to clearly capture the face image under strong light or dim environment, resulting in recognition failure; and wearing hats, glasses, masks and other items may block facial features and affect recognition accuracy. Although technology is constantly improving, under certain conditions (such as twins with high similarity, large changes in facial expressions, changes in appearance caused by aging, etc.), the system may still have misidentification (mistaking different people for the same person) or rejection (unable to correctly identify legitimate users), which will reduce user experience and cause security issues. Face recognition involves the collection, storage and use of personal biometric data. Once these sensitive information is leaked, it may pose a serious threat to user privacy.

[0005] To achieve the above-mentioned object, the present invention aims to provide a smart access control intelligent management system based on face recognition, including a user management unit, which manages and maintains the personal information and access rights of authorized users based on an information management module; A face recognition unit, which captures the facial image of the person entering and leaving the user through the portrait acquisition module based on the user information of the user management unit, introduces the aerosol influence variable for data processing, and performs identity authentication through a convolutional neural network algorithm that introduces the illumination variable; An access control unit, which controls the access control through a control module based on the result of identity authentication of the face recognition unit; The communication integration unit is based on the IP network protocol and establishes a network connection through a wireless network module to ensure stable data transmission between various components in the access control system.

[0006] As a further improvement of the technical solution, the user management unit includes an information management module; The information management module is used to manage user information data and admit new user information data, and to set different access rights for each user according to the user's role.

[0007] As a further improvement of this technical solution, the facial images of people entering and leaving are captured by the portrait acquisition module, the aerosol influence variable is introduced for data processing, and the convolutional neural network algorithm with the illumination variable is introduced for identity authentication. The specific steps involved are as follows: S3.1. When a person approaches the access control system, the portrait acquisition module captures the facial image of the person entering or leaving the access control system. Among them, the facial image includes facial contour, facial features, and skin color; S3.2, introducing the aerosol influence variable to process the facial image captured by the portrait acquisition module to obtain the facial feature vector; S3.3, using the convolutional neural network algorithm with the introduction of illumination variables to analyze the facial feature vector; S3.4, comparing the analysis result with the facial data of the authorized user stored in the user management unit to determine whether there is a match; The authorized user facial data includes the recorded facial image and the recorded user information data; S3.5. Regardless of the verification result, the face recognition unit will record the verification event for subsequent query and audit.

[0008] As a further improvement of the technical solution, in S3.1, the portrait acquisition module includes an infrared sensor module and a camera module; Wherein, the infrared sensor module is used to determine whether there is a person approaching the access control; The camera module is used to capture facial images when the infrared sensor module is triggered.

[0009] As a further improvement of the technical solution, in S3.2, the specific steps of introducing the aerosol influence variable to process the facial image captured by the portrait acquisition module to obtain the facial feature vector are as follows: Input facial image: ; in, represents the input facial image; Indicates the height of the facial image; Indicates the width of the facial image; Indicates the number of channels; represents the set of real numbers; Normalize the facial image: ; in, represents the normalized facial image; represents the mean of facial images; represents the standard deviation of facial images; The expression for correcting the posture by the least squares method is: ; in, Represents attitude parameters; represents the function that transforms key points from the world coordinate system to the image coordinate system; Indicates the position of the reference key point; Indicates the number of key points; Represents the key point index variable; Represents the coordinates of the key point in the world coordinate system; According to the estimated posture parameters , apply the affine transformation matrix Correct the face image to a frontal perspective; The transformed image is: ; in, represents the transformed facial image; represents the affine transformation matrix; Considering that the aerosol influence variable has an impact on the accuracy of the facial feature vector, the aerosol influence variable is introduced to optimize the facial feature vector; Among them, the aerosol-affecting variables include aerosol concentration; Using the Aerosol Concentration Estimation Network After correction, the image Evaluate the degree of its influence by aerosol and obtain the aerosol concentration value; ; in, Indicates the concentration of aerosol; According to the concentration of aerosol Dynamically adjust feature maps , generate the compensated feature map ; ; in, Represents the parameters of the compensation network; represents a fully connected network; The compensated feature map Convert to a fixed-length vector ; ; in, Represents the high-dimensional features of the image; Indicates that the compensated feature map is at position The values ​​of all channels at ; Indicates all channels; Indicates the height of the feature map; Indicates the width of the feature map; A variable representing the height index of the feature map; The width index variable representing the feature map; Normalize the feature vector and output the final feature vector: ; in, Represents the high-dimensional features of the image after normalization; express The Euclidean norm of .

[0010] As a further improvement of the technical solution, in S3.3, the specific process of analyzing the facial feature vector using the convolutional neural network algorithm that introduces the illumination variable is as follows: Using the final feature vector from S3.2 ; Define the Euclidean distance similarity metric: ; in, Represents the feature vector and The Euclidean distance between Represents element index variable; represents the dimension of the feature vector; represents the Euclidean distance; The loss function of the final feature vector is obtained: ; in, represents the loss function of the final feature vector; represents the sample label; Indicates boundary value; Considering that the illumination variable will affect the robustness of the feature vector, the illumination variable is introduced Optimize Among them, the illumination variable Including light intensity and light direction ; ; For lighting variables To perform lighting compensation: ; in, Represents the feature vector after illumination compensation; Represents the parameters of the illumination compensation network; A fully connected network representing illumination compensation; Definition of Euclidean distance similarity metric after introducing illumination variables: ; in, Represents the feature vector and The Euclidean distance between The loss function of the feature vector after illumination compensation is obtained: ; in, Represents the loss function of the feature vector after illumination compensation.

[0011] As a further improvement of the technical solution, the specific steps involved in controlling the access control through the control module are: S7.1. Receive the result of identity authentication from the face recognition unit; S7.2, when the verification result is passed, the control module sends an opening signal to the access control hardware to allow personnel to pass; S7.3, when the verification fails, the access control is kept closed through the control module; S7.4. After the access control action is completed, a confirmation message is sent to the face recognition unit.

[0012] As a further improvement of the technical solution, the control module includes a verification module and a switch module; The verification module is used to make a judgment based on the verification result and then convey instructions to the switch module; The switch module is used to control the switch of the access control.

[0013] As a further improvement of the technical solution, the specific steps involved in establishing a network connection through a wireless network module based on an IP network protocol to ensure stable data transmission between components in the access control system are as follows: S9.1. Use IP network protocol to establish network connection with other components of the access control system through the wireless network module to ensure stable data transmission; S9.2, package, encrypt and send the data exchanged between the units, and receive, decrypt and parse the data packets from other units; S9.3. Monitor the data of each unit and keep them synchronized; S9.4. Monitor the status of network connections, detect and resolve network failures, and ensure the normal operation of the system.

[0014] As a further improvement of the technical solution, in S9.1, the wireless network module includes a wireless communication module and a network interface module; Wherein, the wireless communication module is used for sending and receiving wireless signals; The network interface module is used to manage network connections and process data packets in the network protocol stack.

[0015] Compared with the prior art, the present invention has the following beneficial effects: 1. In this intelligent access control management system based on face recognition, strict permission settings are used to ensure that only authorized personnel can enter specific areas, thereby enhancing the security of the system. Under complex environmental conditions, high recognition accuracy can be maintained, reducing the risk of misidentification. Real-time capture and processing of facial images ensures a fast identity authentication process and improves traffic efficiency. The convolutional neural network algorithm that introduces illumination variables compensates for the influence of illumination, ensuring that the system can stably and accurately identify users under different lighting conditions.

[0016] 2. In this intelligent access control management system based on face recognition, the switch status of the access control is strictly controlled to prevent unauthorized personnel from entering, thus enhancing the security of the system. The use of wireless network connection reduces the wiring cost and installation complexity, and enhances the flexibility and scalability of the system. The IP network protocol ensures reliable data transmission, supports a wide range of network coverage, and is suitable for places of all sizes. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is the overall flow chart of the present invention; The meaning of each number in the figure is: 1. User management unit; 11. Information management module; 2. Face recognition unit; 21. Portrait acquisition module; 211. Infrared sensor module; 212. Camera module; 3. Access control unit; 31. Control module; 311. Verification module; 312. Switch module; 4. Communication integration unit; 41. Wireless network module; 411. Wireless communication module; 412. Network interface module. DETAILED DESCRIPTION

[0018] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0019] See also Figure 1 As shown, a smart access control intelligent management system based on face recognition is provided, including a user management unit 1, which manages and maintains personal information and access rights of authorized users based on an information management module 11; In this example, the user management unit 1 includes an information management module 11; The information management module 11 is used to manage user information data and admit new user information data, and to set different access rights for each user according to the user's role.

[0020] Specifically, the information management module 11 in the user management unit 1 plays a vital role. It is responsible for managing and maintaining detailed information data of all users, including but not limited to personal identity information, facial image data, etc., and supports the entry of new user information. More importantly, the information management module 11 can set appropriate access rights for each user according to the roles of different users (such as residents, visitors, staff, etc.), ensuring that only authorized personnel can enter specific areas, thereby improving the security and management efficiency of the system. In this way, the information management module 11 not only simplifies the management process of user information, but also effectively improves the intelligence level and security of the access control system.

[0021] The smart access control intelligent management system based on face recognition also includes a face recognition unit 2. The face recognition unit 2 captures the facial image of the person entering and leaving the user through the portrait acquisition module 21 based on the user information of the user management unit 1, introduces the aerosol influence variable for data processing, and performs identity authentication through the convolutional neural network algorithm that introduces the illumination variable; In this example, the facial images of people entering and leaving are captured by the portrait acquisition module 21, the aerosol influence variable is introduced for data processing, and the convolutional neural network algorithm with the illumination variable is introduced for identity authentication. The specific steps involved are: S3.1. When a person approaches the access control, the portrait acquisition module 21 captures the facial image of the person entering or leaving the access control; Among them, the facial image includes facial contour, facial features, and skin color; In this example, the portrait acquisition module 21 includes an infrared sensor module 211 and a camera module 212; The infrared sensor module 211 is used to determine whether a person is approaching the access control; The camera module 212 is used to capture facial images when the infrared sensor module 211 is triggered.

[0022] Specifically, the infrared sensor module 211 and the camera module 212 in the portrait acquisition module 21 work together to ensure that the system can efficiently and accurately identify people. The infrared sensor module 211 first detects whether there is a person approaching the access control. Once it senses that someone is approaching, it will immediately trigger the camera module 212 to start, so that the camera begins to prepare to capture facial images. This design not only reduces unnecessary resource consumption (such as electricity and processing power), but also improves the response speed and accuracy of the system, because the camera module is only activated when it is really needed. In this way, the system can quickly prepare for face recognition before the person arrives at the access control, thereby achieving a seamless and secure access management experience.

[0023] S3.2, introducing the aerosol influence variable to process the facial image captured by the portrait acquisition module 21 to obtain a facial feature vector; In this example, the specific steps of introducing the aerosol influence variable to process the facial image captured by the portrait acquisition module 21 and obtaining the facial feature vector are as follows: Input facial image: ; in, represents the input facial image; Indicates the height of the facial image; Indicates the width of the facial image; Indicates the number of channels; represents the set of real numbers; Normalize the facial image: ; in, represents the normalized facial image; represents the mean of facial images; represents the standard deviation of facial images; The expression for correcting the posture by the least squares method is: ; in, Represents attitude parameters; represents the function that transforms key points from the world coordinate system to the image coordinate system; Indicates the position of the reference key point; Indicates the number of key points; Represents the key point index variable; Represents the coordinates of the key point in the world coordinate system; Specifically, The function expression is: ; in, Indicates the key point Coordinates on the axis; Indicates the key point Coordinates on the axis; Indicates the key point Coordinates on the axis; According to the estimated posture parameters , apply the affine transformation matrix Correct the face image to a frontal perspective; The transformed image is: ; in, represents the transformed facial image; represents the affine transformation matrix; Specifically, the expression of the affine transformation matrix is: ; in, Indicates control zoom; Indicates control of rotation; and Indicates controlled shearing; Indicated in The amount of translation in the axis direction; Indicated in The amount of translation in the axis direction; Specifically, this process significantly enhances the robustness and accuracy of face recognition in the smart access control system, especially when dealing with faces in different postures and angles. Through precise posture correction, the system can effectively reduce misidentification caused by changes in head posture, ensuring that only authorized personnel can pass through the access control smoothly, while improving the user experience and the overall security of the system. In addition, the standardized front view also provides a more consistent basis for subsequent feature extraction and matching, further improving recognition efficiency and accuracy.

[0024] Considering that the aerosol influence variable has an impact on the accuracy of the facial feature vector, the aerosol influence variable is introduced to optimize the facial feature vector; Among them, the aerosol-affecting variables include aerosol concentration; Using the Aerosol Concentration Estimation Network After correction, the image Evaluate the degree of its influence by aerosol and obtain the aerosol concentration value; ; in, Indicates the concentration of aerosol; Specifically, the aerosol concentration estimation network for: ; in, Indicates the number of layers in the network; Indicates The weight matrix of the layer; Indicates The weight matrix of the layer; Indicates The bias vector of the layer; Indicates The bias vector of the layer; Indicates The weight matrix of the layer; Indicates The bias vector of the layer; Represents the activation function: in, for: ; in, represents the Euler number; represents the input vector; According to the concentration of aerosol Dynamically adjust feature maps , generate the compensated feature map ; ; in, Represents the parameters of the compensation network; represents a fully connected network; The compensated feature map Convert to a fixed-length vector ; ; in, Represents the high-dimensional features of the image; Indicates that the compensated feature map is at position The values ​​of all channels at ; Indicates all channels; Indicates the height of the feature map; Indicates the width of the feature map; A variable representing the height index of the feature map; The width index variable representing the feature map; Normalize the feature vector and output the final feature vector: ; in, Represents the high-dimensional features of the image after normalization; express The Euclidean norm of .

[0025] Specifically, by accurately evaluating the impact of aerosol and performing corresponding feature compensation, the system can maintain efficient recognition performance in adverse environments, ensuring that only authorized personnel can enter specific areas, thereby improving the overall security of the system and user experience. At the same time, this method also reduces the risk of misidentification, making access management more intelligent and efficient.

[0026] The role of introducing aerosol influence variables is to significantly improve the recognition accuracy and reliability of smart access control management systems based on face recognition in adverse weather conditions (such as fog and haze). By quantifying the impact of aerosol on image quality and dynamically adjusting the feature extraction process to compensate for these adverse factors, the system can more accurately identify facial features and reduce the risk of misidentification caused by environmental interference. This mechanism ensures that the access control system can operate stably even in low visibility conditions, maintaining efficient and safe access management, and enhancing the overall robustness of the system and user experience.

[0027] S3.3, using the convolutional neural network algorithm with the introduction of illumination variables to analyze the facial feature vector; In this example, the specific process of analyzing the facial feature vector using the convolutional neural network algorithm with the introduction of illumination variables is as follows: Using the final feature vector from S3.2 ; Define the Euclidean distance similarity metric: ; in, Represents the feature vector and The Euclidean distance between Represents element index variable; represents the dimension of the feature vector; represents the Euclidean distance; The loss function of the final feature vector is obtained: ; in, represents the loss function of the final feature vector; represents the sample label; Indicates boundary value; Considering that the illumination variable will affect the robustness of the feature vector, the illumination variable is introduced Optimize Among them, the illumination variable Including light intensity and light direction ; ; For lighting variables To perform lighting compensation: ; in, Represents the feature vector after illumination compensation; Represents the parameters of the illumination compensation network; A fully connected network representing illumination compensation; Definition of Euclidean distance similarity metric after introducing illumination variables: ; in, Represents the feature vector and The Euclidean distance between The loss function of the feature vector after illumination compensation is obtained: ; in, Represents the loss function of the feature vector after illumination compensation.

[0028] Specifically, in the intelligent access control intelligent management system based on face recognition, the introduction of the convolutional neural network algorithm with illumination variables to analyze the facial feature vector can significantly improve the robustness and accuracy of the system. By performing illumination compensation on the final feature vector received from the S3.2 module, the system can maintain stable recognition performance under different lighting conditions (such as day, night, indoors, outdoors, etc.). Specifically, the illumination compensation mechanism reduces the impact of illumination changes on the feature vector by estimating and adjusting the illumination intensity and illumination direction, making the feature representation of the face of the same identity more consistent under different lighting conditions. This not only improves the similarity between positive sample pairs, but also increases the difference between negative sample pairs, thereby enhancing the system's ability to distinguish. Ultimately, the loss function after illumination compensation ensures that the model can identify users more accurately, reduces the false recognition rate and rejection rate, and improves user experience and security.

[0029] S3.4, comparing the analysis result with the facial data of the authorized user stored in the user management unit 1 to determine whether there is a match; The authorized user facial data includes the recorded facial image and the recorded user information data; S3.5. Regardless of the verification result, the face recognition unit 2 will record the verification event for subsequent query and audit.

[0030] Specifically, the facial images of people entering and leaving are captured by the portrait acquisition module 21, and the identity authentication process is performed in combination with the convolutional neural network algorithm that takes into account the variables affected by aerosol, ensuring that the system can accurately identify users under various environmental conditions. Specifically, when a person approaches the access control, the system automatically captures their facial image, and uses an optimized deep learning algorithm to analyze the image and extract facial feature points, which can maintain a high recognition accuracy even in the presence of adverse environmental factors such as aerosol. Subsequently, the extracted features are compared with the facial data of authorized users in the database to quickly determine whether the person has access rights. Regardless of the verification result, the system will record the event and provide detailed logs for subsequent inquiries and audits. This process not only enhances the security and reliability of the access control system, but also improves the user experience and ensures efficient and intelligent access management.

[0031] The smart access control intelligent management system based on face recognition also includes an access control unit 3, which controls the access control through a control module 31 based on the result of identity authentication of the face recognition unit 2; In this example, the specific steps involved in controlling the access control through the control module 31 are: S7.1, receiving the identity authentication result from the face recognition unit 2; S7.2, when the verification result is passed, the control module 31 sends an opening signal to the access control hardware to allow the person to pass; In this example, the control module 31 includes a verification module 311 and a switch module 312; The verification module 311 is used to make a judgment based on the verification result and then transmit an instruction to the switch module 312; The switch module 312 is used to control the switch of the access control.

[0032] Specifically, in the smart access control intelligent management system based on face recognition, the verification module 311 and the switch module 312 in the control module 31 work together to ensure the security and automation of the access control operation. Specifically, the verification module 311 makes a judgment based on the identity authentication result provided by the face recognition unit, decides whether to allow passage, and conveys the corresponding instructions to the switch module 312. The switch module 312 accurately controls the opening or closing of the access control according to the received instructions. This mechanism not only realizes the seamless connection between fast and accurate identity authentication and access control, but also ensures that only authorized personnel can enter the controlled area, thereby enhancing the security and reliability of the system, while providing an efficient access management experience.

[0033] S7.3, when the verification fails, the access control is kept closed by the control module 31; S7.4. After the access control action is completed, a confirmation message is sent to the face recognition unit 2.

[0034] Specifically, the control module 31 plays a key bridging role, connecting the identity authentication process with the operation of the physical access control. The control module 31 receives the identity authentication result from the face recognition unit 2, and decides to send an opening signal or keep the access control closed based on whether the verification is passed. When the verification is successful, it triggers the access control hardware to open, allowing authorized personnel to pass; when the verification fails, it ensures that the access control remains locked to prevent unauthorized access. In addition, after each access control action is completed, the control module 31 will also feedback a confirmation message to the face recognition unit 2 to complete the entire control cycle and record the event, ensuring the security of the system and the traceability of the operation. This mechanism is crucial to achieving efficient, safe and automated access management, ensuring that only verified personnel can enter the controlled area.

[0035] The smart access control intelligent management system based on face recognition also includes a communication integration unit 4. The communication integration unit 4 is based on the IP network protocol and establishes a network connection through a wireless network module 41 to ensure stable data transmission between various components in the access control system.

[0036] In this example, based on the IP network protocol, a network connection is established through the wireless network module 41 to ensure stable data transmission between the components in the access control system. The specific steps involved are: S9.1. Use the IP network protocol to establish a network connection with other components of the access control system through the wireless network module 41 to ensure stable data transmission; In this example, the wireless network module 41 includes a wireless communication module 411 and a network interface module 412; Wherein, the wireless communication module 411 is used for sending and receiving wireless signals; The network interface module 412 is used to manage network connections and process data packets in the network protocol stack.

[0037] Specifically, the wireless communication module 411 and the network interface module 412 in the wireless network module 41 jointly ensure efficient, secure and stable wireless data transmission between the components in the system. The wireless communication module 411 is responsible for sending and receiving wireless signals, so that different parts of the access control system can communicate without physical connections, enhancing the flexibility and deployment convenience of the system. The network interface module 412 manages the entire network connection and processes data packets in the network protocol stack, including data packaging, encryption, transmission, and decryption and parsing after reception, to ensure the security and integrity of data transmission. The two work together to not only support real-time data exchange, but also monitor and maintain the health of the network, detect and solve potential network problems in a timely manner, thereby ensuring the stable operation and efficient management of the access control system. This mechanism is crucial to achieving seamless integration and remote management of intelligent access control systems.

[0038] S9.2, package, encrypt and send the data exchanged between the units, and receive, decrypt and parse the data packets from other units; S9.3. Monitor the data of each unit and keep them synchronized; S9.4. Monitor the status of network connections, detect and resolve network failures, and ensure the normal operation of the system.

[0039] Specifically, the stable data transmission mechanism established through the IP network protocol and the wireless network module 41 ensures efficient and secure data interaction between the components in the system. Specifically, this mechanism not only uses the IP network protocol to establish a stable wireless connection to ensure reliable data transmission, but also packages and encrypts the exchanged data to ensure data security and privacy. At the same time, it continuously monitors the data synchronization status between the units and monitors the network connection status in real time. It can detect and resolve possible network failures in a timely manner, thereby maintaining the normal operation and response speed of the system. This process is crucial to realizing the intelligent management of the access control system and providing uninterrupted security protection, ensuring the smooth operation of each link from user information management to face recognition verification.

[0040] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments, and the above embodiments and descriptions are only preferred examples of the present invention, and are not intended to limit the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, and these changes and improvements all fall within the scope of the present invention to be protected.

Claims

1. Intelligent access control management system based on face recognition, characterized by: include A user management unit (1), the user management unit (1) managing and maintaining personal information and access rights of authorized users based on the information management module (11); A face recognition unit (2), wherein the face recognition unit (2) captures a facial image of a person entering or leaving the user through a portrait acquisition module (21) based on the user information of the user management unit (1), introduces an aerosol influence variable for data processing, and performs identity authentication through a convolutional neural network algorithm that introduces an illumination variable; An access control unit (3), the access control unit (3) controlling the access control through a control module (31) based on the identity verification result of the face recognition unit (2); A communication integration unit (4), the communication integration unit (4) establishing a network connection through a wireless network module (41) based on an IP network protocol, thereby ensuring stable data transmission between various components in the access control system.

2. The smart access control intelligent management system based on face recognition according to claim 1 is characterized by: The user management unit (1) comprises an information management module (11); The information management module (11) is used to manage user information data and admit new user information data, and to set different access rights for each user according to the user's role.

3. The smart access control intelligent management system based on face recognition according to claim 1 is characterized by: The face recognition unit (2) captures the facial image of the person entering or leaving through the portrait acquisition module (21), introduces the aerosol influence variable for data processing, and performs identity authentication through the convolutional neural network algorithm that introduces the illumination variable. The specific steps involved are: S3.

1. When a person approaches the access control system, the portrait acquisition module (21) captures the facial image of the person entering or leaving the access control system; Among them, the facial image includes facial contour, facial features, and skin color; S3.2, introducing the aerosol influence variable to process the facial image captured by the portrait acquisition module (21) to obtain a facial feature vector; S3.3, using the convolutional neural network algorithm with the introduction of illumination variables to analyze the facial feature vector; S3.4, comparing the analysis result with the facial data of the authorized user stored in the user management unit (1) to determine whether there is a match; The authorized user facial data includes the recorded facial image and the recorded user information data; S3.

5. Regardless of the verification result, the face recognition unit (2) will record the verification event for subsequent query and audit.

4. The smart access control intelligent management system based on face recognition according to claim 3 is characterized by: In S3.1, the portrait acquisition module (21) includes an infrared sensor module (211) and a camera module (212); The infrared sensor module (211) is used to determine whether a person is approaching the access control system; The camera module (212) is used to capture a facial image of a person when the infrared sensor module (211) is triggered.

5. The smart access control intelligent management system based on face recognition according to claim 3 is characterized by: In S3.2, the specific steps of introducing the aerosol influence variable to process the facial image captured by the portrait acquisition module (21) to obtain the facial feature vector are: Input facial image: ； in, represents the input facial image; Indicates the height of the facial image; Indicates the width of the facial image; Indicates the number of channels; represents the set of real numbers; Normalize the facial image: ； in, represents the normalized facial image; represents the mean of facial images; represents the standard deviation of facial images; The expression for correcting the posture by the least squares method is: ； in, Represents attitude parameters; represents the function that transforms key points from the world coordinate system to the image coordinate system; Indicates the position of the reference key point; Indicates the number of key points; Represents the key point index variable; Represents the coordinates of the key point in the world coordinate system; According to the estimated posture parameters , apply the affine transformation matrix Correct the face image to a frontal perspective; The transformed image is: ； in, represents the transformed facial image; represents the affine transformation matrix; Considering that the aerosol influence variable has an impact on the accuracy of the facial feature vector, the aerosol influence variable is introduced to optimize the facial feature vector; Among them, the aerosol-affecting variables include aerosol concentration; Using the Aerosol Concentration Estimation Network After correction, the image Assess the degree of its impact by aerosol and obtain the aerosol concentration value; ； in, Indicates the concentration of aerosol; According to the concentration of aerosol Dynamically adjust feature maps , generate the compensated feature map ; ； in, Represents the parameters of the compensation network; represents a fully connected network; The compensated feature map Convert to a fixed-length vector ; ； in, Represents the high-dimensional features of the image; Indicates that the compensated feature map is at position The values ​​of all channels at ; Indicates all channels; Indicates the height of the feature map; Indicates the width of the feature map; A variable representing the height index of the feature map; The width index variable representing the feature map; Normalize the feature vector and output the final feature vector: ； in, Represents the high-dimensional features of the image after normalization; express The Euclidean norm of .

6. The smart access control intelligent management system based on face recognition according to claim 5 is characterized by: In S3.3, the specific process of analyzing the facial feature vector using the convolutional neural network algorithm that introduces the illumination variable is as follows: Using the final feature vector from S3.2 ; Define the Euclidean distance similarity metric: ； in, Represents the feature vector and The Euclidean distance between Represents element index variable; represents the dimension of the feature vector; represents the Euclidean distance; The loss function of the final feature vector is obtained: ； in, represents the loss function of the final feature vector; represents the sample label; Indicates boundary value; Considering that the illumination variable will affect the robustness of the feature vector, the illumination variable is introduced Optimize Among them, the illumination variable Including light intensity and light direction ; ; For lighting variables To perform lighting compensation: ; in, represents the feature vector after illumination compensation; Represents the parameters of the illumination compensation network; A fully connected network representing illumination compensation; Definition of Euclidean distance similarity metric after introducing illumination variables: ； in, Represents the feature vector and The Euclidean distance between The loss function of the feature vector after illumination compensation is obtained: ； in, Represents the loss function of the feature vector after illumination compensation.

7. The smart access control intelligent management system based on face recognition according to claim 1 is characterized by: The specific steps involved in the access control unit (3) controlling the access control via the control module (31) are: S7.

1. Receive the identity verification result from the face recognition unit (2); S7.2, when the verification result is passed, the control module (31) sends an opening signal to the access control hardware to allow the person to pass; S7.3, when the verification fails, the access control is kept closed through the control module (31); S7.

4. After the access control action is completed, a confirmation message is sent to the face recognition unit (2).

8. The smart access control intelligent management system based on face recognition according to claim 7 is characterized by: The control module (31) comprises a verification module (311) and a switch module (312); The verification module (311) is used to make a judgment based on the verification result and then transmit an instruction to the switch module (312); The switch module (312) is used to control the switch of the access control.

9. The smart access control intelligent management system based on face recognition according to claim 1 is characterized by: The communication integration unit (4) establishes a network connection through a wireless network module (41) based on an IP network protocol to ensure stable data transmission between components in the access control system. The specific steps involved are: S9.

1. Use the IP network protocol to establish a network connection with other components of the access control system through the wireless network module (41) to ensure stable data transmission; S9.2, package, encrypt and send the data exchanged between the units, and receive, decrypt and parse the data packets from other units; S9.

3. Monitor the data of each unit and keep them synchronized; S9.

4. Monitor the status of network connections, detect and resolve network failures, and ensure the normal operation of the system.

10. The smart access control intelligent management system based on face recognition according to claim 9 is characterized by: In S9.1, the wireless network module (41) includes a wireless communication module (411) and a network interface module (412); Wherein, the wireless communication module (411) is used for sending and receiving wireless signals; The network interface module (412) is used to manage network connections and process data packets in the network protocol stack.

Images