Access control equipment with AI face recognition function

By combining visible light and infrared camera image acquisition in the access control device, complementing face features and adopting low-power wake-up mode, the problem of poor face recognition effect under strong and low-light conditions is solved, and the recognition accuracy is improved and the convenience of access control management.

CN222050954UActive Publication Date: 2024-11-22广州万境信息技术有限公司
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202420975137.8
Authority / Receiving Office
CN · China
Patent Type
Utility models(China)
Current Assignee / Owner
Filing Date
2024-05-08
Publication Date
2024-11-22
Estimated Expiration
2034-05-08

AI Technical Summary

Technical Problem

Existing AI-based access control equipment has poor facial recognition effect under strong and low-light conditions, especially in low-light environments at night, which affects the recognition rate.

Method used

The visible light camera and infrared camera are used to participate in the face image acquisition at the same time. The infrared image and the visible light image are complementary, and combined with the low-power wake-up mode and the unit gate state monitoring mode, AI face recognition is realized through NPU chips and logic gates.

Benefits of technology

The accuracy and image quality of face recognition under different lighting conditions are improved, circuit power consumption is reduced, convenient access control management experience is achieved, and the unit door status is timely monitored and reported.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN222050954U_ABST
    Figure CN222050954U_ABST
Patent Text Reader

Abstract

The utility model discloses access control equipment with an AI face recognition function. The access control equipment comprises a unit door machine, an indoor terminal and a property management terminal, the unit door machine comprises a host and a lock driving module which are arranged on a door plate of the unit door, and a door magnetic sensor arranged on the door plate of the unit door; the host comprises a shell, an electronic circuit board and a low-power-consumption wake-up circuit, the electronic circuit board and the low-power-consumption wake-up circuit are arranged in the shell, a first NPU chip, a second NPU chip, a logic gate and a CPU chip are integrated on the electronic circuit board, a visible light camera and an infrared camera are arranged in the shell, and holes for exposing the visible light camera and the infrared camera are formed in the shell. The AI face recognition device is simple in structure and reasonable in design, the visible light camera and the infrared camera participate in face image acquisition at the same time, infrared images and visible light images are complementary, the AI face recognition device is used for achieving AI face recognition, and the AI face recognition device has a low-power-consumption awakening mode and a cell gate state monitoring mode.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The utility model belongs to the technical field of security, and specifically relates to an access control device with an AI face recognition function. Background Art

[0002] The access control system based on AI recognition breaks the single and mechanical button unlocking method of traditional video intercom, reshapes the control interaction at various distances, and realizes a convenient and user-friendly access control management experience. However, the current access control system based on AI recognition mostly uses visible light images for the collection of facial images. Image recognition at night mostly uses fill light, which is mostly used to illuminate the front of the face. If the fill light is not in place, the facial image will have both strong light and weak light images. The image quality of hardware imaging under strong light and weak light conditions is essentially different. There is too much noise in weak light images, which seriously affects the recognition rate.

[0003] The study found that the images collected by visible light cameras have clear contours, rich details, and are three-channel RGB images, which can effectively reflect facial features. However, the quality of visible light images is greatly affected by lighting conditions. The intensity of light in the actual environment is often uncontrollable, especially at night, which makes it difficult to identify facial features in low-light environments. Infrared cameras can achieve all-weather and all-day imaging, and have the advantages of being insensitive to light intensity and strong penetration. Even in the case of rain, fog, haze and other obstructions, the impact on infrared images is relatively low. The disadvantages of infrared images include: blurred edges, unclear picture details, single-channel images, only brightness information, sensitive to changes in ambient temperature, and non-penetrating glass. Through the above comparison of advantages and disadvantages, it is found that infrared images and visible light images have complementary characteristics. Images obtained under visible light and images obtained under near-infrared light reflect the characteristics of images in two different modes.

[0004] Moreover, most traditional access control devices are on standby all day long, so a low-power access control device based on visible light and infrared to achieve AI face recognition is needed. Utility Model Content

[0005] The technical problem to be solved by the utility model is to provide an access control device with AI face recognition function in view of the deficiencies in the above-mentioned prior art. The device has a simple structure and a reasonable design. The visible light camera and the infrared camera are simultaneously involved in the face image acquisition. The infrared image and the visible light image complement each other and are used to realize AI face recognition. The device also has a low-power wake-up mode and a unit door status monitoring mode.

[0006] In order to solve the above technical problems, the technical solution adopted by the utility model is: an access control device with AI face recognition function, characterized in that: it includes a unit door machine arranged on the unit door, an indoor terminal arranged in the owner's home, and a property terminal arranged in the property office, which can communicate with each other; the unit door machine includes a host arranged on the door panel of the unit door, a lock drive module, and a door magnetic sensor arranged on the door of the unit door; the host includes a shell, and an electronic circuit board and a low-power wake-up circuit arranged inside the shell, and the electronic circuit board integrates a first NPU chip, a second NPU chip, a logic gate and a CPU chip A piece of film is provided in a shell, a visible light camera and an infrared camera are arranged in the shell, a hole is arranged on the shell, and the visible light camera and the infrared camera are exposed from the holes respectively; the low-power wake-up circuit includes an infrared sensor, a comparator U1 and an LDO power supply module arranged in the shell, a shutter that can emit infrared rays is arranged on the shell, the output end of the infrared sensor is connected to the reverse input end of the comparator U1, the comparison voltage is connected to the same-direction input end of the comparator U1, the output end of the comparator U1 is connected to the enable end of the LDO power supply module, and the output end of the LDO power supply module supplies power to various power-consuming modules on the electronic circuit board.

[0007] The above-mentioned access control device with AI face recognition function is characterized in that: the visible light camera is connected to the input end of the first NPU chip, the infrared camera is connected to the input end of the second NPU chip, the output end of the first NPU chip and the output end of the second NPU chip are respectively connected to the input end of the logic gate, the output end of the logic gate is connected to the input end of the CPU chip, and the output end of the CPU chip is connected to the input end of the lock drive module.

[0008] The above-mentioned access control device with AI face recognition function is characterized in that: the first NPU chip is connected to the first image database to obtain the first target feature set; the first NPU chip determines whether the face image collected by the visible light camera belongs to the first target feature set; the second NPU chip is connected to the second image database to obtain the second target feature set; the second NPU chip determines whether the face image collected by the infrared camera belongs to the second target feature set.

[0009] The above-mentioned access control device with AI face recognition function is characterized in that the logic gate is an OR gate.

[0010] The above-mentioned access control device with AI face recognition function is characterized in that: the door magnetic sensor includes a door magnetic main part installed on the door panel through a first mounting frame and a door magnetic auxiliary part installed on the door frame through a second mounting frame, the door magnetic sensor also includes a microprocessor and a wireless communication unit connected to the microprocessor, and the wireless communication unit communicates wirelessly with the CPU chip; a first conductive sheet is arranged on the door magnetic main part, and a second conductive sheet is arranged on the door magnetic auxiliary part, the first conductive sheet and the second conductive sheet constitute a section of wire in a resistor voltage divider circuit, and the voltage divider output end of the resistor voltage divider circuit is connected to the microprocessor.

[0011] The above-mentioned access control device with AI face recognition function is characterized in that: the ends of the first conductive sheet and the second conductive sheet that are in contact are both semicircular arc-shaped, the first conductive sheet and the second conductive sheet are arranged facing each other, a first self-restoring spring is arranged between the first conductive sheet and the door magnetic main component, and a second self-restoring spring is arranged between the second conductive sheet and the door magnetic auxiliary component.

[0012] The above-mentioned access control device with AI face recognition function is characterized in that: the power supply provides the working voltage for the LDO power supply module, and the comparison voltage is the bias voltage of the power supply.

[0013] The above-mentioned access control device with AI face recognition function is characterized in that a keyboard for password unlocking is connected to the host, and the keyboard is exposed from the shell.

[0014] The above-mentioned access control device with AI face recognition function is characterized in that a card reader for swiping a card to unlock the door is connected to the host.

[0015] Compared with the prior art, the utility model has the following advantages:

[0016] 1. The utility model has a simple structure, reasonable design, and is easy to implement and operate.

[0017] 2. In the present invention, the visible light camera and the infrared camera are simultaneously involved in the facial image acquisition, and the infrared image and the visible light image complement each other to realize AI facial recognition.

[0018] 3. In the utility model, the infrared sensor and the LDO power supply module form a low-power wake-up circuit, which reduces the power consumption of the circuit, realizes ultra-low power standby, and has a good use effect.

[0019] 4. The utility model adopts a door magnetic sensor to monitor the switch status of the unit door, and can report to the property terminal to promptly remind the property and prevent accidents before they happen.

[0020] To sum up, the utility model has a simple structure and a reasonable design. The visible light camera and the infrared camera are simultaneously involved in facial image acquisition. The infrared image and the visible light image complement each other and are used to realize AI face recognition. It also has a low-power wake-up mode and a unit door status monitoring mode.

[0021] The technical solution of the utility model is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 This is a circuit principle block diagram of the utility model.

[0023] Figure 2 This is a schematic diagram of the installation of the unit door machine of the utility model.

[0024] Figure 3 This is a schematic diagram of the installation of the door magnetic sensor of the utility model.

[0025] Figure 4 The utility model is a circuit principle diagram of the resistor voltage divider circuit.

[0026] Figure 5 It is a structural schematic diagram of the host of the utility model.

[0027] Figure 6 This is a circuit principle block diagram of the host of the utility model.

[0028] Figure 7 This is a circuit schematic diagram of the low-power wake-up circuit of the utility model.

[0029] Description of reference numerals:

[0030] 1—unit door machine; 2—indoor terminal; 3—property terminal;

[0031] 11—CPU chip; 12—logic gate; 13—first NPU chip;

[0032] 14—second NPU chip; 15—LDO power supply module; 16—power supply;

[0033] 17—infrared sensor; 18—visible light camera; 19—first image database;

[0034] 110—infrared camera; 111—second image database; 112—microphone;

[0035] 113—door magnetic sensor; 114—housing; 115—louver;

[0036] 116—display screen; 117—lock drive module; 1131—first mounting frame;

[0037] 1132—Door magnet main part; 1133—Door magnet auxiliary part; 1134—Second mounting frame;

[0038] 1135—first conductive sheet; 1136—second conductive sheet; 4—host; DETAILED DESCRIPTION

[0039] The present invention will be further described in detail below with reference to the accompanying drawings and embodiments of the present invention.

[0040] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present invention will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0041] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present application. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprise" and / or "include" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof.

[0042] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein, for example. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0043] For ease of description, spatially relative terms such as "above", "above", "on the upper surface of", "above", etc. may be used here to describe the spatial positional relationship between a device or feature and other devices or features as shown in the figure. It should be understood that spatially relative terms are intended to include different orientations of the device in use or operation in addition to the orientation described in the figure. For example, if the device in the accompanying drawings is inverted, the device described as "above other devices or structures" or "above other devices or structures" will be positioned as "below other devices or structures" or "below other devices or structures". Thus, the exemplary term "above" can include both "above" and "below". The device can also be positioned in other different ways (rotated 90 degrees or in other orientations), and the spatially relative descriptions used here are interpreted accordingly.

[0044] like Figure 1 As shown, the utility model includes a unit door machine 1 arranged on the unit door, an indoor terminal 2 arranged in the owner's home, and a property terminal 3 arranged in the property office, which can communicate with each other. In actual use, the number of unit door machines 1 and indoor terminals 2 is multiple. The unit door machine 1 and the indoor terminal 2 are connected by bus. The unit door machine 1 and the property terminal 3 communicate using TCP / IP protocol. The indoor terminal 2 is connected to the local area network through an RJ45 network cable to achieve communication with the property terminal 3.

[0045] The unit door machine 1 comprises a host 4 arranged on the door panel of the unit door, a lock driving module 117, and a door magnetic sensor 113 arranged on the door of the unit door. The host 4 is embedded in the door panel of the unit door.

[0046] The host 4 is electrically connected to the lock driving module 117 and the door magnetic sensor 113 respectively. The lock driving module 117 is used to open the door lock according to the host instruction, and the door magnetic sensor 113 is used to report the open or closed state of the door to the host 4.

[0047] The host 4 includes a shell 114, and an electronic circuit board and a low-power wake-up circuit arranged inside the shell 114. The electronic circuit board integrates a first NPU chip 13, a second NPU chip 14, a logic gate 12 and a CPU chip 11. A visible light camera 18 and an infrared camera 110 are arranged in the shell 114. Holes are arranged on the shell 114, and the visible light camera 18 and the infrared camera 110 are respectively exposed from the holes.

[0048] In this embodiment, the visible light camera 18 is connected to the input end of the first NPU chip 13, the infrared camera 110 is connected to the input end of the second NPU chip 14, the output end of the first NPU chip 13 and the output end of the second NPU chip 14 are respectively connected to the input end of the logic gate 12, the output end of the logic gate 12 is connected to the input end of the CPU chip 11, and the output end of the CPU chip 11 is connected to the input end of the lock driving module 117. The CPU chip 11 is also connected to a display screen 116 and a microphone 112.

[0049] The images captured by the visible light camera 18 have clear contours, rich detail information, and are three-channel RGB images, which can effectively reflect facial features. However, the quality of visible light images is greatly affected by lighting conditions. The intensity of light in the actual environment is often uncontrollable, especially at night, which makes it difficult to identify facial features in low-light environments.

[0050] The infrared camera 110 can realize all-weather and all-day imaging, and has the advantages of being insensitive to light intensity and strong penetration. Even in the presence of obstructions such as rain, fog, and haze, the impact on infrared images is relatively low. The disadvantages of infrared images include: blurred edges, unclear picture details, single-channel images with only brightness information, sensitive to changes in ambient temperature, and non-penetrating to glass.

[0051] Through the above comparison of advantages and disadvantages, it is found that infrared images and visible light images have complementary characteristics. The images obtained under visible light and the images obtained under near infrared reflect the characteristics of the images in two different modes, so the visible light camera 18 and the infrared camera 110 are involved in the face image acquisition at the same time.

[0052] The visible light camera 18 captures images of human faces under visible light, and sends the visible light images to the first NPU chip 13. The first NPU chip 13 is connected to the first image database 19 to obtain a first target feature set. The first NPU chip 13 determines whether the visible light image captured by the visible light camera 18 belongs to the first target feature set, and sends the determination result to the logic gate 12.

[0053] The infrared camera 110 collects an infrared image of a human face and sends the infrared image to the second NPU chip 14. The second NPU chip 14 is connected to the second image database 111 to obtain a second target feature set. The second NPU chip 14 determines whether the facial image collected by the infrared camera 110 belongs to the second target feature set and sends the determination result to the logic gate 12.

[0054] In this embodiment, the logic gate 12 is an OR gate, that is, if the first NPU chip 13 passes the judgment, or the second NPU chip 14 passes the judgment, the OR gate output is high; if and only if the judgments of both the first NPU chip 13 and the second NPU chip 14 fail, the OR gate output is low.

[0055] When the CPU chip 11 receives the high level output of the logic gate 12, the CPU chip 11 drives the lock driving module 117 to work, and the lock driving module 117 opens the door lock. In a possible embodiment, the door lock is an electromagnetic lock, and the lock driving module 117 is used to drive the electromagnetic lock to unlock when receiving the control signal of the CPU chip 11.

[0056] The low-power wake-up circuit includes an infrared sensor 17, a comparator U1 and an LDO power supply module 15 arranged in a shell 114. A shutter 115 is arranged on the shell 114 to emit infrared rays. The output end of the infrared sensor 17 is connected to the reverse input end of the comparator U1, the comparison voltage is connected to the same-direction input end of the comparator U1, the output end of the comparator U1 is connected to the enable end of the LDO power supply module 15, and the output end of the LDO power supply module 15 supplies power to various power-consuming modules on the electronic circuit board.

[0057] In actual use, the infrared sensor 17 uses a chip of model HC-SR501. The output end of the chip HC-SR501 is connected to the in-phase input end of the comparator U1. The LHI778 probe of the chip HC-SR501 continuously emits infrared rays outside the shell 114 through the shutter 15. When the emitted infrared rays are not reflected or the intensity of the reflected rays is not strong enough, the output end of the chip HC-SR501 is a low level, 0V; when the detected human body appears in the detection range, the infrared rays are reflected back and the intensity is strong enough. At this time, the output end of the chip HC-SR501 is a high level, 3.3V.

[0058] Therefore, the comparison voltage is set to 1.5V. When the output end of the chip HC-SR501 is at a high level of 3.3V, the output of the comparator U1 is high, so the enable end of the LDO power supply module 15 is at a high level, the LDO power supply module 15 starts working, exits the low-power working mode, and supplies power to various power-consuming modules on the electronic circuit board; when the output end of the chip HC-SR501 is at a low level of 0V, the output of the comparator U1 is low, the enable end of the LDO power supply module 15 is at a low level, and the LDO power supply module 15 does not work.

[0059] In this embodiment, the door magnetic sensor 113 includes a door magnetic main component 1132 installed on the door panel through a first mounting frame 1131 and a door magnetic sub-component 1133 installed on the door frame through a second mounting frame 1134. The door magnetic sensor 113 also includes a microprocessor and a wireless communication unit connected to the microprocessor, and the wireless communication unit wirelessly communicates with the CPU chip 11; a first conductive sheet is arranged on the door magnetic main component 1132, and a second conductive sheet is arranged on the door magnetic sub-component 1133. The first conductive sheet and the second conductive sheet constitute a section of wire in a resistor voltage divider circuit, and the voltage divider output end of the resistor voltage divider circuit is connected to the microprocessor.

[0060] Specifically, the resistor voltage divider circuit includes a first wire, a first conductive sheet, a second conductive sheet, a second wire, a resistor R1, a resistor R2 and a third wire connected in sequence, one end of the first wire is connected to a power supply, and one end of the third wire is grounded.

[0061] The first mounting bracket 1131 and the second mounting bracket 1134 are provided to ensure the position of the door magnetic sensor 113, and the door magnetic main component 1132 and the door magnetic auxiliary component 1133 will not be knocked off by abnormal situations. When the door is in an open state, one end of the first conductive sheet is separated from one end of the second conductive sheet, so the resistor voltage divider circuit is not turned on, and the voltage divider output end of the resistor voltage divider circuit is 0V, that is, low level. The microprocessor transmits the door opening time to the CPU chip 11 through the wireless communication unit for timely prompting the property. The wireless communication unit includes but is not limited to ZigBee, Bluetooth, 433 and other transmission methods; when the door is in a closed state, one end of the first conductive sheet and one end of the second conductive sheet are closed, the resistor voltage divider circuit is turned on, the voltage divider output end of the resistor voltage divider circuit is high level, and the microprocessor does not operate.

[0062] In this embodiment, the contact ends of the first conductive sheet and the second conductive sheet are both semicircular, the first conductive sheet and the second conductive sheet are arranged facing each other, a first self-restoring spring is arranged between the first conductive sheet and the door magnetic main component 1132, and a second self-restoring spring is arranged between the second conductive sheet and the door magnetic auxiliary component 1133.

[0063] The semicircular arc design increases the probability of the first conductive sheet and the second conductive sheet touching and fitting each other. When the door is opened, the first conductive sheet returns to the semicircular arc shape under the restoring force of the first self-restoring spring, and the second conductive sheet returns to the semicircular arc shape under the restoring force of the second self-restoring spring.

[0064] The microprocessor is a single-chip microcomputer MSP430F169, and the voltage-dividing output end of the resistor voltage-dividing circuit is connected to the P1.0 pin of the single-chip microcomputer MSP430F169.

[0065] The first NPU chip 13 and the second NPU chip 14 both use Rising Micro RV1126 chips.

[0066] The power supply 16 provides the working voltage for the LDO power supply module 15, and the comparison voltage is the bias voltage of the power supply 16. In a possible embodiment, the LDO power supply module 15 uses the chip NCP730.

[0067] In actual use, the host 4 is connected with a keyboard for password unlocking, and the keyboard is exposed from the housing 114. The host 4 is connected with a card reader for swiping a card to unlock.

[0068] Among them, the contents not described in detail in the specification belong to the existing technology known to professional and technical personnel in this field.

[0069] The above description is only an embodiment of the present invention and does not constitute any limitation to the present invention. Any simple modification, change and equivalent structural change made to the above embodiments according to the technical essence of the present invention shall still fall within the protection scope of the technical solution of the present invention.

Claims

1. An access control device with AI face recognition function, characterized by: It comprises a unit door machine (1) arranged on a unit door and capable of communicating with each other, an indoor terminal (2) arranged in the owner's home, and a property terminal (3) arranged in a property office; The unit door machine (1) comprises a host (4) arranged on a door panel of the unit door, a lock drive module (117), and a door magnetic sensor (113) arranged on the door of the unit door; The host (4) comprises a housing (114), and an electronic circuit board and a low-power wake-up circuit arranged inside the housing (114); a first NPU chip (13), a second NPU chip (14), a logic gate (12) and a CPU chip (11) are integrated on the electronic circuit board; a visible light camera (18) and an infrared camera (110) are arranged inside the housing (114); a hole is arranged on the housing (114), and the visible light camera (18) and the infrared camera (110) are respectively exposed from the hole; The low-power wake-up circuit comprises an infrared sensor (17), a comparator U1 and an LDO power supply module (15) arranged in a housing (114); a shutter (115) capable of emitting infrared rays is arranged on the housing (114); the output end of the infrared sensor (17) is connected to the reverse input end of the comparator U1, the comparison voltage is connected to the same-direction input end of the comparator U1, the output end of the comparator U1 is connected to the enable end of the LDO power supply module (15), and the output end of the LDO power supply module (15) supplies power to various power-consuming modules on the electronic circuit board.

2. The access control device with AI face recognition function according to claim 1, characterized in that: The visible light camera (18) is connected to the input end of the first NPU chip (13), the infrared camera (110) is connected to the input end of the second NPU chip (14), the output end of the first NPU chip (13) and the output end of the second NPU chip (14) are respectively connected to the input end of the logic gate (12), the output end of the logic gate (12) is connected to the input end of the CPU chip (11), and the output end of the CPU chip (11) is connected to the input end of the lock drive module (117).

3. An access control device with AI face recognition function according to claim 1 or 2, characterized in that: The first NPU chip (13) is connected to the first image database (19) to obtain a first target feature set; the first NPU chip (13) determines whether the face image collected by the visible light camera (18) belongs to the first target feature set; the second NPU chip (14) is connected to the second image database (111) to obtain a second target feature set; the second NPU chip (14) determines whether the face image collected by the infrared camera (110) belongs to the second target feature set.

4. The access control device with AI face recognition function according to claim 2, characterized in that: The logic gate (12) is an OR gate.

5. The access control device with AI face recognition function according to claim 1, characterized in that: The door magnetic sensor (113) comprises a door magnetic main component (1132) mounted on the door panel via a first mounting frame (1131) and a door magnetic auxiliary component (1133) mounted on the door frame via a second mounting frame (1134). The door magnetic sensor (113) further comprises a microprocessor and a wireless communication unit connected to the microprocessor, and the wireless communication unit communicates wirelessly with the CPU chip (11); A first conductive sheet is arranged on the door magnetic main part (1132), and a second conductive sheet is arranged on the door magnetic auxiliary part (1133). The first conductive sheet and the second conductive sheet form a section of a conductor in a resistor voltage divider circuit, and a voltage divider output end of the resistor voltage divider circuit is connected to a microprocessor.

6. The access control device with AI face recognition function according to claim 5, characterized in that: The ends of the first conductive sheet and the second conductive sheet that are in contact are both semicircular arc-shaped. The first conductive sheet and the second conductive sheet are arranged facing each other. A first self-restoring spring is arranged between the first conductive sheet and the door magnetic main component (1132), and a second self-restoring spring is arranged between the second conductive sheet and the door magnetic auxiliary component (1133).

7. The access control device with AI face recognition function according to claim 1, characterized in that: The power supply (16) provides an operating voltage for the LDO power supply module (15), and the comparison voltage is a bias voltage of the power supply (16).

8. The access control device with AI face recognition function according to claim 1, characterized in that: The host (4) is connected to a keyboard for password unlocking, and the keyboard is exposed from the housing (114).

9. The access control device with AI face recognition function according to claim 1, characterized in that: The host (4) is connected with a card reader for unlocking by swiping a card.