Intelligent bus charging system based on face recognition and control method thereof

Abstract

The invention relates to the technical field of intelligent bus charging systems based on face recognition, in particular to an intelligent bus charging system based on face recognition and a control method thereof.The intelligent bus charging system based on face recognition comprises a face recognition module, an image information processing module, a background deduction module and a payment confirmation module; the background deduction module comprises a personal information management module and a cost management module; the face recognition module is used for realizing face recognition, extracting face digital information and storing the face digital information into a database; according to the invention, aiming at the unique features of the face of the elderly, the multi-scale information input module collects the feature information of the face of the elderly, designs a polynomial loss adjustment strategy to adjust the weight, guides the elderly to extract or pay attention to the key features of the face, adds a feature enhancement attention module to enhance the key features, and adapts to the loss strategy. And challenges caused by the complexity of the facial features of the old people are solved in a targeted manner.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent public transportation fare collection systems based on face recognition, and in particular to an intelligent public transportation fare collection system based on face recognition and a control method thereof. Background Art

[0002] The intelligent bus fare collection system with face recognition is a modern system that uses biometric technology to realize public transportation payment. The system captures the facial image of the passenger through the camera installed on the bus, and then analyzes and compares it through advanced face recognition algorithms to identify and verify the passenger's identity. Once the identity is confirmed, the system will automatically deduct the corresponding fare from the passenger's account to complete the payment process. This system not only improves the convenience and efficiency of bus payment, but also reduces the dependence on traditional ticket staff. At the same time, it also provides passengers with a payment method without carrying cash or cards;

[0003] At the same time, the system can also collect and analyze passenger flow data to provide support for urban traffic management and planning. With the continuous advancement of technology, this system is expected to play a more important role in future intelligent transportation systems.

[0004] The existing face recognition system proposes different face recognition algorithms for specific scenarios, including the image blur problem in motion scenes, the face fatigue state recognition system, etc. Although the face recognition algorithm for specific scenarios is more accurate than the universal algorithm in actual application, in the process of face recognition applied to elderly scenes, due to the wrinkles, eyelid bags and other features of the elderly face, it is difficult to achieve the expected effect when a large number of algorithms are applied in the intelligent public transportation fare collection system based on face recognition. Therefore, in order to solve the above problems, an intelligent public transportation fare collection system based on face recognition and its control method are proposed. Summary of the invention

[0005] The purpose of the present invention is to provide an intelligent public transportation fare collection system based on face recognition and a control method thereof, so as to solve the problem that the existing face recognition system proposes different face recognition algorithms for specific scenes, taking into account the image blur problem in motion scenes, the face fatigue state recognition system, etc. Although the face recognition algorithm for specific scenes is more accurate than the universal algorithm in actual application, in the process of face recognition applied to elderly scenes, due to the presence of wrinkles, eyelid bags and other features on the elderly's face, a large number of algorithms are applied in the intelligent public transportation fare collection system based on face recognition, and it is difficult to achieve the expected effect.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] An intelligent public transportation charging system based on face recognition and a control method thereof, comprising a face recognition module, an image information processing module, a background deduction module and a payment confirmation module, wherein the face recognition module comprises a face image acquisition module, and the background deduction module comprises a personal information management module and a fee management module; the face recognition module is used to realize face recognition, extract digital face information, and store it in a database; the face image acquisition module is used to automatically capture a face image when a passenger approaches the front door of a bus and aligns his head with the front of a face device; the image information processing module is used to obtain the captured face image and process it; the background deduction module is used to receive processed face image information data and execute a deduction operation, the personal information management module is used to add, delete, modify and query personal information, and the fee management module is used to manage user balances and user fee details; the payment confirmation module is used to transmit the deduction result to a front-end device and display payment success information on a client.

[0008] As a further optimized content of the present invention, the facial image acquisition module includes at least one camera, and the camera is configured with automatic focus and automatic exposure functions.

[0009] As a further optimized content of the present invention, the image information processing module includes face detection, feature extraction and similarity comparison algorithms, which are used to compare the acquired facial image with the facial images stored in the database.

[0010] As a further optimized content of the present invention, the background deduction module includes account verification, transaction processing and security encryption algorithms to ensure the security and accuracy of transactions.

[0011] As a further optimized content of the present invention, the payment confirmation module includes a real-time communication protocol and a user interface, which is used to display the payment result to the passenger in real time and provide a prompt of payment success or failure.

[0012] As a further optimized content of the present invention, the system further includes a user registration module for collecting and registering the passenger's facial image and account information when used for the first time.

[0013] As a further optimization of the present invention, the system further includes an exception handling module for providing alternative payment options or prompts when facial recognition fails or payment is abnormal.

[0014] As a further optimized content of the present invention, the following steps are included:

[0015] Step a: Capture facial images. When a passenger approaches the front door of the bus and points his head to the front of the facial device, the camera automatically captures the facial image and performs facial recognition through the facial recognition module. In the process of facial recognition, the features are segmented through the age-based sub-classifier, the elderly are classified into age groups, and the facial features of each age group are specially optimized;

[0016] Step b: Obtain image information. After the passenger confirms the upload, the system processes it and transmits the image for similarity comparison and is processed by the image information processing module;

[0017] Step c: Backstage deduction operation, transmit facial information data to the backstage, and perform deduction operation;

[0018] Step d: Confirm the payment information. The backend transmits the relevant results to the front-end device. The front-end device receives the information and displays it to the client, and reminds that the payment is successful.

[0019] As a further optimization of the present invention, when the face recognition module recognizes that the face is an elderly person, S1: extracts face features through a multi-scale information input module, uses convolution units of different sizes, and fuses features through Concat, the formula is:

[0020]

[0021] Where: f1, f2 and f3 represent the feature maps output by the convolution unit, H and W are the height and width of the feature map, Ci is the number of convolution channels, Bi, Ti and Ei are adjustment factors;

[0022] S2: Extract features using average pooling and convolution units through the global adaptive feature extraction module, and smooth the features using ReLU and Sigmoid activation functions;

[0023] S3: Combined with the feature enhancement attention module, pixel attention mechanism and maximum pooling are used to enhance the attention to the unique facial features of the elderly;

[0024] S4: Adopt adaptive loss strategy to optimize recognition results;

[0025] As a further optimized content of the present invention, the adaptive loss strategy adopts an improved EIoU loss function, and the formula is:

[0026]

[0027] IoU is the intersection over union ratio, which is used to measure the degree of overlap between the predicted box and the true box. D(b, b_true) is the distance between the predicted box and the true box. c is the diagonal length of the minimum bounding box that can contain both the predicted box and the true box. w and w_n are the widths of the predicted box and the true box, h and h_n are the heights of the predicted box and the true box, U and L are the width and height of the minimum bounding box respectively. A and L are weight coefficients. R1, R2 and R3 are constants used to adjust the various indicators in the loss function.

[0028] Compared with the prior art, the present invention has the following beneficial effects:

[0029] 1. In the present invention, in view of the unique features of the elderly face, the multi-scale information input module collects the feature information of the elderly face, and designs a polynomial loss adjustment strategy to adjust the weight, guides it to extract or focus on the key features of the face, adds a feature enhancement attention module to enhance the key features, and adopts adaptive loss strategy and other technical means to solve the challenges brought by the complexity of the facial features of the elderly in a targeted manner;

[0030] 2. In the present invention, the global and local features of the face are extracted by using convolution units of different sizes, thereby enhancing the system's ability to recognize facial features of the elderly. At the same time, average pooling and convolution units are used to extract features, and ReLU and Sigmoid activation functions are used to smooth features, thereby reducing the interference of features such as wrinkles and improving recognition accuracy. In addition, the pixel attention mechanism and maximum pooling are combined to enhance attention to the unique facial features of the elderly, further improving recognition accuracy.

[0031] 3. In the present invention, an improved E IoU loss function is adopted to optimize the training process of the model, so that the system can recognize the faces of the elderly more accurately and facilitate the registration of new users, ensuring that only registered users can use the facial recognition payment function, thereby enhancing the security of the system. At the same time, a backup payment option is provided when facial recognition fails or payment is abnormal, thereby ensuring the continuity of the transaction, improving the user experience, and ensuring that the payment results can be quickly fed back to the passengers, providing timely payment success or failure prompts. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 This is a flow chart of an intelligent public transportation fare collection system based on face recognition and a control method thereof according to the present invention;

[0033] Figure 2 The present invention is a flow chart of the intelligent public transportation fare collection system based on face recognition. DETAILED DESCRIPTION

[0034] See also Figure 1-2 , the present invention provides a technical solution:

[0035] The intelligent public transportation charging system based on face recognition and its control method include a face recognition module, an image information processing module, a background deduction module and a payment confirmation module, wherein the face recognition module includes a face image acquisition module, and the background deduction module includes a personal information management module and a fee management module; the face recognition module is used to realize face recognition, extract digital face information, and store it in a database; the face image acquisition module is used to automatically capture a face image when a passenger approaches the front door of a bus and aligns his head with the front of a face device; the image information processing module is used to obtain the captured face image and process it; the background deduction module is used to receive processed face image information data and execute a deduction operation, the personal information management module is used to add, delete, modify and query personal information, and the fee management module is used to manage user balances and user fee details; the payment confirmation module is used to transmit the deduction result to a front-end device and display payment success information on a client. The system realizes the automation and intelligence of public transportation charging through the integrated face recognition module and the background deduction module, improves the operation efficiency of the public transportation system, and provides passengers with a convenient payment experience. The introduction of personal information management and expense management modules has enhanced the system's user management and financial management capabilities, ensuring the security and accuracy of transactions;

[0036] As a technical solution for further implementation of the present invention, the facial image acquisition module includes at least one camera, which is configured with auto-focus and auto-exposure functions. By configuring the camera with auto-focus and auto-exposure functions, it is ensured that clear and high-quality facial images can be captured even under different lighting conditions and distances, thereby improving the accuracy of face recognition and the reliability of the system;

[0037] As a technical solution for further implementation of this solution, the image information processing module includes face detection, feature extraction and similarity comparison algorithms, which are used to compare the acquired face image with the face image stored in the database. The face detection, feature extraction and similarity comparison algorithms of the image information processing module enable the system to effectively process and analyze the face image, ensuring the efficiency and accuracy of face recognition, while improving the response speed of the system and user experience;

[0038] As a technical solution for further implementation of this solution, the backend deduction module includes account verification, transaction processing and security encryption algorithms to ensure the security and accuracy of transactions. The account verification, transaction processing and security encryption algorithms of the backend deduction module provide strong security protection for the public transportation fare collection system, ensure the security and accuracy of transactions, enhance users' trust in the system, and also protect users' financial information;

[0039] As a technical solution for further implementation of this solution, the payment confirmation module includes a real-time communication protocol and a user interface, which is used to display the payment result to the passenger in real time and provide a prompt of payment success or failure. The real-time communication protocol and user interface of the payment confirmation module ensure that the payment result can be quickly fed back to the passenger, provide timely payment success or failure prompts, and improve user satisfaction and system interaction experience;

[0040] As a technical solution for further implementation of this solution, the system further includes a user registration module for collecting and registering the passenger's facial image and account information when used for the first time. The introduction of the user registration module enables new users to easily register and use the face recognition payment function, while also providing the system with a mechanism for managing user accounts and facial images, thereby enhancing the system's user management and service capabilities;

[0041] As a technical solution for further implementation of this solution, the system also includes an exception handling module, which is used to provide alternative payment options or prompts when facial recognition fails or payment is abnormal. The design of the exception handling module ensures that when facial recognition fails or an abnormality occurs during the payment process, the system can provide effective solutions, such as alternative payment options, thereby ensuring the continuity of transactions and the user's payment experience;

[0042] As a technical solution for further implementing this plan, the following steps are included:

[0043] Step a: Capture facial images. When a passenger approaches the front door of the bus and points his head to the front of the facial device, the camera automatically captures the facial image and performs facial recognition through the facial recognition module. In the process of facial recognition, the features are segmented through the age-based sub-classifier, the elderly are classified into age groups, and the facial features of each age group are specially optimized;

[0044] Step b: Obtain image information. After the passenger confirms the upload, the system processes it and transmits the image for similarity comparison and is processed by the image information processing module;

[0045] Step c: Backstage deduction operation, transmit facial information data to the backstage, and perform deduction operation;

[0046] Step d: Confirm the payment information. The backend transmits the relevant results to the front-end device. The front-end device receives the information and displays it to the client, and reminds the client that the payment is successful. The efficiency and accuracy of public transportation fees are improved through automated processes and optimized face recognition technology. In particular, when dealing with facial features of the elderly, the recognition accuracy is improved through special optimization based on age-based sub-classifiers, ensuring that all user groups can enjoy convenient payment services.

[0047] As a technical solution for further implementation of this solution, when the face recognition module recognizes that the face is an elderly person, S1: extracts face features through the multi-scale information input module, uses convolution units of different sizes, and fuses features through Concat, the formula is:

[0048]

[0049] Where: f1, f2 and f3 represent the feature maps output by the convolution unit, H and W are the height and width of the feature map, Ci is the number of convolution channels, Bi, Ti and Ei are adjustment factors;

[0050] S2: Extract features using average pooling and convolution units through the global adaptive feature extraction module, and smooth the features using ReLU and Sigmoid activation functions;

[0051] S3: Combined with the feature enhancement attention module, pixel attention mechanism and maximum pooling are used to enhance the attention to the unique facial features of the elderly;

[0052] S4: Adopt adaptive loss strategy to optimize recognition results;

[0053] The adaptive loss strategy adopts the improved EIoU loss function, the formula is:

[0054]

[0055] IoU is the intersection over union ratio, which is used to measure the degree of overlap between the predicted box and the true box. D(b,b_true) is the distance between the predicted box and the true box. c is the diagonal length of the minimum bounding box that can contain both the predicted box and the true box. w and w_n are the widths of the predicted box and the true box, h and h_n are the heights of the predicted box and the true box, U and L are the width and height of the minimum bounding box respectively. A and L are weight coefficients. R1, R2 and R3 are constants used to adjust the various indicators in the loss function. In the facial recognition scenario of the elderly, the challenges brought by the complexity of the facial features of the elderly are specifically addressed by introducing technical means such as feature enhancement attention mechanism and adaptive loss strategy.

[0056] The present invention uses specific examples to illustrate the principles and implementation methods of the present invention. The above examples are only used to help understand the methods and core ideas of the present invention. The above is only a preferred implementation of the present invention. It should be pointed out that due to the limitations of textual expression, there are objectively infinite specific structures. For ordinary technicians in this technical field, without departing from the principles of the present invention, they can make several improvements, modifications or changes, and can also combine the above technical features in an appropriate manner; these improvements, modifications, changes or combinations, or the direct application of the inventive concept and technical solution to other occasions without improvement, should be regarded as the protection scope of the present invention.

Claims

1. Intelligent public transportation fare collection system based on face recognition, characterized in that: It includes a face recognition module, an image information processing module, a backend deduction module and a payment confirmation module, wherein the face recognition module includes a facial image acquisition module, and the backend deduction module includes a personal information management module and a fee management module; The face recognition module is used to realize face recognition, extract digital face information, and store it in a database; The facial image acquisition module is used to automatically capture a facial image when a passenger approaches the front door of the bus and aligns his head with the front of the facial device; The image information processing module is used to obtain and process the photographed facial image; The backend deduction module is used to receive the processed facial image information data and perform the deduction operation, the personal information management module is used to add, delete, modify and query personal information, and the fee management module is used to manage the user balance and user fee details; The payment confirmation module is used to transmit the deduction result to the front-end device and display the payment success information on the client.

2. The intelligent public transportation fare collection system based on face recognition according to claim 1 is characterized in that: The facial image acquisition module includes at least one camera, which is equipped with automatic focus and automatic exposure functions.

3. The intelligent public transportation fare collection system based on face recognition according to claim 1 is characterized in that: The image information processing module includes face detection, feature extraction and similarity comparison algorithms, which are used to compare the acquired face image with the face images stored in the database.

4. The intelligent public transportation fare collection system based on face recognition according to claim 1 is characterized in that: The backend deduction module includes account verification, transaction processing and security encryption algorithms to ensure the security and accuracy of transactions.

5. The intelligent public transportation fare collection system based on face recognition according to claim 1 is characterized in that: The payment confirmation module includes a real-time communication protocol and a user interface, which is used to display the payment result to the passenger in real time and provide a prompt of payment success or failure.

6. The intelligent public transportation fare collection system based on face recognition according to claim 1 is characterized in that: The system further includes a user registration module for collecting and registering a passenger's facial image and account information upon first use.

7. The intelligent public transportation fare collection system based on face recognition according to claim 1 is characterized in that: The system also includes an exception handling module for providing alternative payment options or prompts when facial recognition fails or payment is abnormal.

8. The control method of the intelligent public transportation fare collection system based on face recognition according to claim 1 is characterized in that: The following steps are involved: Step a: Capture facial images. When a passenger approaches the front door of the bus and points his head to the front of the facial device, the camera automatically captures the facial image and performs facial recognition through the facial recognition module. In the process of facial recognition, the features are segmented through the age-based sub-classifier, the elderly are classified into age groups, and the facial features of each age group are specially optimized; Step b: Obtain image information. After the passenger confirms the upload, the system processes it and transmits the image for similarity comparison and is processed by the image information processing module; Step c: Backstage deduction operation, transmit facial information data to the backstage, and perform deduction operation; Step d: Confirm the payment information. The backend transmits the relevant results to the front-end device. The front-end device receives the information and displays it to the client, and reminds that the payment is successful.

9. The control method of the intelligent public transportation fare collection system based on face recognition according to claim 8 is characterized in that: When the face recognition module identifies the face as an elderly person, S1: extracts face features through the multi-scale information input module, uses convolution units of different sizes, and fuses features through Concat, the formula is: Where: f1, f2 and f3 represent the feature maps output by the convolution unit, H and W are the height and width of the feature map, Ci is the number of convolution channels, Bi, Ti and Ei are adjustment factors; S2: Extract features using average pooling and convolution units through the global adaptive feature extraction module, and smooth the features using ReLU and Sigmoid activation functions; S3: Combined with the feature enhancement attention module, pixel attention mechanism and maximum pooling are used to enhance the attention to the unique facial features of the elderly; S4: Adopt an adaptive loss strategy to optimize the recognition results.

10. The control method of the intelligent public transportation fare collection system based on face recognition according to claim 9 is characterized in that: The adaptive loss strategy adopts an improved EIoU loss function, the formula is: IoU is the intersection over union ratio, which is used to measure the degree of overlap between the predicted box and the true box. D(b, b_true) is the distance between the predicted box and the true box. c is the diagonal length of the minimum bounding box that can contain both the predicted box and the true box. w and w_n are the widths of the predicted box and the true box, h and h_n are the heights of the predicted box and the true box, U and L are the width and height of the minimum bounding box respectively. A and L are weight coefficients. R1, R2 and R3 are constants used to adjust the various indicators in the loss function.

Images