An electronic class board for face recognition based on a TPU acceleration board

Abstract

The utility model relates to electronic class board system technical field, concretely, a kind of electronic class board for face recognition based on TPU acceleration board;Through the setting camera head gathers face image and is transmitted to TPU acceleration development board;According to the face image obtained from camera head, generate face recognition result by setting TPU acceleration development board;According to the face recognition result obtained from TPU acceleration development board, generation unlocking instruction by setting raspberry pi 4B development board;According to the unlocking instruction obtained from raspberry pi 4B development board, execute corresponding action by setting relay module, complete network-independent fast face recognition, and satisfy the urgent needs of school scene for low power consumption, real-time processing and offline operation.

Description

Technical Field

[0001] This utility model relates to the field of electronic class sign system technology, specifically to an electronic class sign for face recognition based on a TPU accelerator board. Background Technology

[0002] With the rapid development of smart campuses, electronic class signs have gradually become common information technology devices in schools. Placed at classroom entrances or in public areas, they are primarily used to display class announcements, timetables, and campus culture promotions. They can also be integrated with multimedia interactive functions to achieve teacher and student information management and home-school collaboration. However, to further enhance the convenience and security of campus management, electronic class signs have incorporated access control and attendance functions. This involves using a camera to capture the facial images of students or teachers for facial recognition and to complete access control or attendance records.

[0003] In traditional technologies, facial recognition systems often use cloud servers for identification. This approach requires uploading facial images captured by a camera to a remote server via the network for processing and comparison, with the recognition result then returned to the local terminal. While this method can functionally achieve facial verification, it still has significant shortcomings in the following aspects: Network Dependence and Latency: Cloud-based recognition is highly dependent on the network environment. When the network is unstable or bandwidth is insufficient, recognition efficiency drops sharply, even affecting the normal passage of students and faculty. Furthermore, any network latency can degrade the user experience, preventing the smooth "instant recognition" effect from being achieved.

[0004] Privacy and security risks: Uploading biometric data such as facial features to the cloud carries the risk of data leakage or interception. Privacy protection is especially important for students; unauthorized access to facial information could lead to privacy breaches and potential security threats.

[0005] High costs: Building or renting cloud servers requires additional hardware and maintenance costs. If the scale of use by a school is large, even more budget needs to be invested in upgrading bandwidth and server resources. At the same time, the high concurrency processing of cloud servers also requires continuous operational investment. Utility Model Content

[0006] This invention addresses the issues of network latency, privacy and security, and high costs associated with traditional cloud-based facial recognition methods. It proposes an electronic class sign for facial recognition based on a TPU accelerator board. The system uses a camera to capture facial images and transmits them to the TPU accelerator board. The TPU accelerator board generates facial recognition results based on the images from the camera. A Raspberry Pi 4B development board generates an unlocking command based on the facial recognition results from the TPU accelerator board. A relay module executes the corresponding action based on the unlocking command from the Raspberry Pi 4B development board, achieving fast, network-independent facial recognition and meeting the urgent needs of schools for low power consumption, real-time processing, and offline operation.

[0007] The specific implementation details of this utility model are as follows: An electronic class sign for face recognition based on a TPU accelerator board includes a Raspberry Pi 4B development board, a TPU accelerator board, a relay module, and a camera; The Raspberry Pi 4B development board is connected to the TPU accelerator development board via an Ethernet interface. The Raspberry Pi 4B development board is connected to the display screen via an HDMI interface; The Raspberry Pi 4B development board is connected to the relay module via a GPIO interface; The camera is connected to the CSI interface of the Raspberry Pi 4B development board via the CSI interface. The camera is used to capture facial images and transmit them to the TPU accelerator board; The TPU acceleration development board is used to generate face recognition results based on the face image obtained from the camera; The Raspberry Pi 4B development board is used to generate unlocking commands based on the face recognition results obtained from the TPU accelerator development board. The relay module is used to execute corresponding actions based on the unlocking command obtained from the Raspberry Pi 4B development board.

[0008] To better realize this utility model, the relay module further includes an electromagnetic relay and an electromagnetic lock; The GND pin of the electromagnetic relay is connected to the Pin9 pin of the Raspberry Pi 4B development board, the IN pin of the electromagnetic relay is connected to the Pin11 pin of the Raspberry Pi 4B development board, the COM pin of the electromagnetic relay is connected to the positive power supply, and the ON pin of the electromagnetic relay is connected to the positive terminal of the Raspberry Pi electromagnetic lock. The negative terminal of the electromagnetic lock is connected to the negative terminal of the power supply.

[0009] To better realize this utility model, the electronic class sign for face recognition based on the TPU accelerator board further includes an RFID radio frequency IC card sensing module; the SDA pin of the RFID radio frequency IC card sensing module is connected to the Pin3 pin of the Raspberry Pi 4B development board; the SCL pin of the RFID radio frequency IC card sensing module is connected to the Pin5 pin of the Raspberry Pi 4B development board.

[0010] To better realize this utility model, the electronic class sign for face recognition based on the TPU accelerator board further includes a display screen; The display screen is connected to the display interface of the Raspberry Pi 4B development board via an HDMI interface, and the touch interface of the display screen is connected to the USB interface of the Raspberry Pi 4B development board.

[0011] To better realize this utility model, the TPU acceleration development board is further described as an NVIDIA Jetson Nano Developer Kit acceleration board.

[0012] To better realize this utility model, the electromagnetic relay is further described as an SRD-05VDC-SL-C relay.

[0013] To better realize this utility model, the RFID radio frequency IC card sensing module is further described as an RC522 RFID radio frequency IC card.

[0014] This utility model has the following beneficial effects: (1) This utility model performs inference on TPU, which is fast and can complete a complete face comparison in hundreds of milliseconds or less, meeting the needs of high-frequency face recognition attendance on campus.

[0015] (2) This utility model does not require uploading data to the cloud. The entire recognition process is completed on the electronic class card device. Even if the campus LAN or the Internet is interrupted, it will not affect the normal access control and attendance functions.

[0016] (3) Compared with traditional GPUs, the TPU module in this invention has a higher computing power-to-power ratio; compared with using CPU to force the deep learning model to run, TPU can significantly reduce power consumption under the same computing power, which has great advantages for deploying a large number of class sign devices.

[0017] (4) In this utility model, the facial images and biometric data are stored and processed only on the local device, which greatly reduces the risk of cloud leakage and is more in line with the campus requirements for student privacy and security. Attached Figure Description

[0018] Figure 1A schematic diagram of the electronic class sign structure for face recognition based on a TPU accelerator board provided by this utility model.

[0019] Figure 2 A schematic block diagram of the overall structure of the electronic class sign for face recognition based on a TPU accelerator board provided by this utility model. Detailed Implementation

[0020] To more clearly illustrate the technical solutions of the embodiments of this utility model, the technical solutions of the embodiments of this utility model will be clearly and completely described below with reference to the accompanying drawings. It should be understood that the described embodiments are only some embodiments of this utility model, not all embodiments, and therefore should not be regarded as a limitation on the scope of protection. Based on the embodiments of this utility model, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this utility model.

[0021] In the description of this utility model, it should be noted that, unless otherwise explicitly specified and limited, the terms "set up," "connected," and "linked" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this utility model based on the specific circumstances.

[0022] Example 1: This embodiment proposes an electronic class sign for face recognition based on a TPU accelerator board, including a Raspberry Pi 4B development board, a TPU accelerator board, a relay module, and a camera; The Raspberry Pi 4B development board is connected to the TPU accelerator development board via an Ethernet interface. The Raspberry Pi 4B development board is connected to the display screen via an HDMI interface; The Raspberry Pi 4B development board is connected to the relay module via a GPIO interface; The camera is connected to the CSI interface of the Raspberry Pi 4B development board via the CSI interface. The camera is used to capture facial images and transmit them to the TPU accelerator board; The TPU acceleration development board is used to generate face recognition results based on the face image obtained from the camera; The Raspberry Pi 4B development board is used to generate unlocking commands based on the face recognition results obtained from the TPU accelerator development board. The relay module is used to execute corresponding actions based on the unlocking command obtained from the Raspberry Pi 4B development board.

[0023] Working principle: In this embodiment, the camera preferably uses a USB or MIPI interface to connect to the Raspberry Pi 4B development board. The video stream or image frames captured by the camera are immediately transmitted to the CPU or TPU inference engine of the development board to prepare for subsequent face detection and recognition. The camera position can be fixed according to the scene to ensure a suitable framing range and clarity.

[0024] In this embodiment, access control can be implemented using GPIO, relays, or dry contacts to connect the opening and closing circuit of the electromagnetic lock. The NFC reader reads the card ID and transmits it to the Raspberry Pi 4b development board. After successful verification by the Raspberry Pi 4b development board, a signal command is sent to trigger the relay, automatically opening the door.

[0025] In this embodiment, the Raspberry Pi 4B development board receives the face recognition result returned by the TPU, and triggers the access control unlocking command only when the verification is successful.

[0026] The TPU accelerator board in this embodiment is the NVIDIA Jetson Nano Developer Kit accelerator board.

[0027] Example 2: This embodiment is based on the above embodiment 1, such as... Figure 1 As shown, the relay module includes an electromagnetic relay and an electromagnetic lock; The GND pin of the electromagnetic relay is connected to the Pin9 pin of the Raspberry Pi 4B development board, the IN pin of the electromagnetic relay is connected to the Pin11 pin of the Raspberry Pi 4B development board, the COM pin of the electromagnetic relay is connected to the positive power supply, and the ON pin of the electromagnetic relay is connected to the positive terminal of the Raspberry Pi electromagnetic lock. The negative terminal of the electromagnetic lock is connected to the negative terminal of the power supply.

[0028] Working principle: In this embodiment, the electromagnetic relay is a 5V relay SRD-05VDC-SL-C, and the electromagnetic lock is a 12V DC electric lock MG-12.

[0029] The other parts of this embodiment are the same as those in Embodiment 1 above, so they will not be described again.

[0030] Example 3: This embodiment is based on any one of Embodiments 1-2 above, such as Figure 1 As shown, the electronic class sign for face recognition based on the TPU accelerator board also includes an RFID radio frequency IC card sensing module; the SDA pin of the RFID radio frequency IC card sensing module is connected to the Pin3 pin of the Raspberry Pi 4B development board; the SCL pin of the RFID radio frequency IC card sensing module is connected to the Pin5 pin of the Raspberry Pi 4B development board.

[0031] Working principle: The RFID radio frequency IC card sensing module in this embodiment is an RC522 radio frequency module.

[0032] The other parts of this embodiment are the same as any one of the above embodiments 1-2, so they will not be described again.

[0033] Example 4: This embodiment is based on any one of embodiments 1-3 above, such as Figure 1 As shown, the electronic class sign for face recognition based on a TPU accelerator board also includes a display screen; The display screen is connected to the DISPLAY interface of the Raspberry Pi 4B development board via an HDMI interface.

[0034] Working principle: In this embodiment, the display screen is connected to the graphics card interface (such as HDMI or MIPI DSI) of the development board to output the recognition results. In practical applications, teachers or students can directly see the attendance status on the class sign, such as "Check-in successful" or "Face recognition passed".

[0035] The other parts of this embodiment are the same as any one of the embodiments 1-3 above, so they will not be described again.

[0036] The above description is merely a preferred embodiment of the present utility model and is not intended to limit the present utility model in any way. Any simple modifications or equivalent changes made to the above embodiments based on the technical essence of the present utility model shall fall within the protection scope of the present utility model.

Claims

1. An electronic class sign for face recognition based on a TPU accelerator board, characterized in that, Includes Raspberry Pi 4B development board, TPU accelerator development board, relay module, and camera; The Raspberry Pi 4B development board is connected to the TPU accelerator development board via an Ethernet interface. The Raspberry Pi 4B development board is connected to the display screen via an HDMI interface; The Raspberry Pi 4B development board is connected to the relay module via a GPIO interface; The camera is connected to the CSI interface of the Raspberry Pi 4B development board via the CSI interface. The camera is used to capture facial images and transmit them to the TPU accelerator board; The TPU acceleration development board is used to generate face recognition results based on the face image obtained from the camera; The Raspberry Pi 4B development board is used to generate unlocking commands based on the face recognition results obtained from the TPU accelerator development board. The relay module is used to execute corresponding actions based on the unlocking command obtained from the Raspberry Pi 4B development board.

2. The electronic class sign for face recognition based on a TPU accelerator board according to claim 1, characterized in that, The relay module includes an electromagnetic relay and an electromagnetic lock; The GND pin of the electromagnetic relay is connected to the Pin9 pin of the Raspberry Pi 4B development board, the IN pin of the electromagnetic relay is connected to the Pin11 pin of the Raspberry Pi 4B development board, the COM pin of the electromagnetic relay is connected to the positive power supply, and the ON pin of the electromagnetic relay is connected to the positive terminal of the Raspberry Pi electromagnetic lock. The negative terminal of the electromagnetic lock is connected to the negative terminal of the power supply.

3. The electronic class sign for face recognition based on a TPU accelerator board according to claim 1, characterized in that, The electronic class sign for face recognition based on the TPU accelerator board also includes an RFID radio frequency IC card sensing module; the SDA pin of the RFID radio frequency IC card sensing module is connected to the Pin3 pin of the Raspberry Pi 4B development board; the SCL pin of the RFID radio frequency IC card sensing module is connected to the Pin5 pin of the Raspberry Pi 4B development board.

4. The electronic class sign for face recognition based on a TPU accelerator board according to claim 1, characterized in that, The electronic class sign for face recognition based on the TPU accelerator board also includes a display screen; The HDMI interface of the display screen is connected to the DISPLAY interface of the Raspberry Pi 4B development board; the Touch interface of the display screen is connected to the USB interface of the Raspberry Pi 4B development board.

5. An electronic class sign for face recognition based on a TPU accelerator board according to claim 1, characterized in that, The TPU acceleration development board is the NVIDIA Jetson Nano Developer Kit acceleration board.

6. An electronic class sign for face recognition based on a TPU accelerator board according to claim 2, characterized in that, The electromagnetic relay is an SRD-05VDC-SL-C relay.

7. An electronic class sign for face recognition based on a TPU accelerator board according to claim 3, characterized in that, The RFID radio frequency IC card sensing module is an RC522 RFID radio frequency IC card.

Images