High-precision water stain fingerprint identification method and device

Abstract

The invention discloses a high-precision water stain fingerprint identification method and device, and the method comprises the steps: obtaining a miniature humidity sensor, monitoring and obtaining a humidity value of a fingerprint contact region through the miniature humidity sensor, carrying out the image collection according to the humidity value, and obtaining a first fingerprint feature image; obtaining a deep learning denoising model, training the deep learning denoising model to obtain a target deep learning denoising model, and repairing the first fingerprint feature image by using the target deep learning denoising model to obtain a second fingerprint feature image, finger subcutaneous vein features are extracted to carry out auxiliary verification and secondary restoration on the second fingerprint feature image, and a final fingerprint feature image is obtained; and performing fingerprint matching identification on the final fingerprint feature image. The identification precision of the water stain fingerprints is improved, and the problem that an existing fingerprint identification module is high in fingerprint identification failure rate when water is attached to the fingers is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of water stain fingerprint recognition of a fingerprint module, and in particular to a high-precision water stain fingerprint recognition method and device. Background Art

[0002] The optical fingerprint module uses the principle of light refraction and reflection. Light is emitted from the bottom to the prism and then emitted through the prism. The angle and darkness of the light refraction and reflection on the uneven lines of the fingerprint on the surface of the finger are different. The CMOS or CCD optical device collects image information of different brightness and darkness to complete the fingerprint collection. The fingerprint feature points and fingerprint feature map will then be converted into digital signals, which will be matched with the digital signals in the database to achieve fingerprint recognition.

[0003] In current fingerprint recognition modules, if the user's finger is wet, the moisture will cause light scattering, blurring the fingerprint image, and the water film will interfere with the electric field distribution of the capacitive sensor, making it impossible to accurately read the texture, ultimately leading to recognition failure. Summary of the Invention

[0004] Existing fingerprint recognition modules can easily fail to recognize fingerprints when there is moisture on the finger.

[0005] To address the above problems, a high-precision water-soaked fingerprint recognition method and device are proposed. The method uses a micro humidity sensor to monitor and obtain the humidity value of the fingerprint contact area, performs image acquisition based on the humidity value, obtains a first fingerprint feature image, and uses the target deep learning denoising model to repair the first fingerprint feature image to obtain a second fingerprint feature image. The subcutaneous vein features of the finger are extracted to perform auxiliary verification and secondary repair on the second fingerprint feature image to obtain the final fingerprint feature image. This improves the recognition accuracy of water-soaked fingerprints and solves the problem of high fingerprint recognition failure rate of existing fingerprint recognition modules when the finger is wet.

[0006] In a first aspect, a high-precision water fingerprint recognition method is provided, comprising: Step 100: Obtain a micro humidity sensor, use the micro humidity sensor to monitor and obtain a humidity value of the fingerprint contact area, perform image acquisition based on the humidity value, and obtain a first fingerprint feature image; Step 200: Obtain a deep learning denoising model, train the deep learning denoising model to obtain a target deep learning denoising model, use the target deep learning denoising model to repair the first fingerprint feature image to obtain a second fingerprint feature image, extract the finger subcutaneous vein features to perform auxiliary verification and secondary repair on the second fingerprint feature image to obtain a final fingerprint feature image; Step 300: Perform fingerprint matching and recognition on the final fingerprint feature image.

[0007] In conjunction with the high-precision water fingerprint recognition method described in the first aspect of the present invention, in a first possible implementation, step 100 includes: Step 110: Obtain a micro humidity sensor; Step 120: embed the micro humidity sensor on the surface of the fingerprint recognition area.

[0008] In conjunction with the first possible implementation of the first aspect of the present invention, in a second possible implementation, step 100 further includes: Step 130: Obtain the humidity value; Step 140: If the humidity value is less than a specified threshold, a visible light image acquisition mode is adopted to obtain the first fingerprint feature image.

[0009] In conjunction with the first possible implementation of the first aspect of the present invention, in a third possible implementation, step 100 further includes: Step 150: Obtain the humidity value; Step 160: If the humidity value is greater than a specified threshold, infrared penetration imaging is used to capture a fingerprint image, and an enhanced capacitive sensor is used to perform fingerprint feature scanning to enhance the fingerprint features on the fingerprint image to obtain the first fingerprint feature image.

[0010] In conjunction with the first possible implementation of the first aspect of the present invention, in a fourth possible implementation, step 100 further includes: Step 170: Obtain the humidity value; Step 180: If the humidity value is greater than a specified threshold, the user is prompted to adjust the fingerprint touch position through screen display or vibration.

[0011] In conjunction with the high-precision water fingerprint recognition method described in the first aspect of the present invention, in a fifth possible implementation, step 200 includes: Step 210: Obtain an input layer, multiple convolutional layers, a batch normalization layer, an activation function layer, and an output layer; Step 220: Build a deep learning denoising model using the input layer, multiple convolutional layers, batch normalization layer, activation function layer, and output layer.

[0012] In combination with the fifth possible implementation manner of the first aspect of the present invention, in a sixth possible implementation manner, step 200 further includes: Step 230: Collect and obtain fingerprint image sample data of different noise types, and obtain a mean square error function; Step 240: Input the fingerprint image sample data into the deep learning denoising model for training, and use the mean square error function to evaluate the difference between the predicted denoised fingerprint image and the true clean fingerprint image, and modify the parameters of the deep learning denoising model based on the evaluation result; Step 250: After multiple rounds of training, the target deep learning denoising model is obtained.

[0013] In a second aspect, a high-precision water stain fingerprint recognition device adopts the high-precision water stain fingerprint recognition method described in the first aspect, comprising: An image acquisition module, configured to acquire a micro humidity sensor, monitor and acquire a humidity value of the fingerprint contact area using the micro humidity sensor, and acquire an image based on the humidity value to obtain a first fingerprint feature image, wherein the micro humidity sensor is embedded in the surface of the fingerprint recognition area; a denoising module, configured to obtain a deep learning denoising model, train the deep learning denoising model to obtain a target deep learning denoising model, use the target deep learning denoising model to repair the first fingerprint feature image to obtain a second fingerprint feature image, extract the finger subcutaneous vein features to perform auxiliary verification and secondary repair on the second fingerprint feature image to obtain a final fingerprint feature image; The recognition module is used to perform fingerprint matching recognition on the final fingerprint feature image.

[0014] In conjunction with the high-precision water fingerprint recognition device described in the second aspect of the present invention, in a first possible implementation, the image acquisition module includes: A first image acquisition unit, a second image acquisition unit and a prompt unit; The first image acquisition unit is configured to acquire the humidity value, and when the humidity value is less than a specified threshold, adopt a visible light image acquisition mode to obtain the first fingerprint feature image; The second image acquisition unit is configured to acquire the humidity value, and when the humidity value is greater than a specified threshold, acquire a fingerprint image using infrared penetration imaging, and enhance the fingerprint features on the fingerprint image by performing fingerprint feature scanning using an enhanced capacitive sensor to obtain the first fingerprint feature image; The prompting unit is used to obtain the humidity value, and when the humidity value is greater than a specified threshold, prompt the user through screen display or vibration to adjust the fingerprint touch position.

[0015] In conjunction with the high-precision water fingerprint recognition device described in the second aspect of the present invention, in a second possible implementation, the denoising module includes: A model construction unit, configured to obtain an input layer, multiple convolutional layers, a batch normalization layer, an activation function layer, and an output layer, and construct a deep learning denoising model using the input layer, multiple convolutional layers, batch normalization layers, activation function layers, and output layer; A training unit is used to collect and obtain fingerprint image sample data of different noise types and obtain a mean square error function, input the fingerprint image sample data into the deep learning denoising model for training, and use the mean square error function to evaluate the difference between the predicted denoised fingerprint image and the real clean fingerprint image, and correct the parameters of the deep learning denoising model according to the evaluation result. After multiple rounds of training, the target deep learning denoising model is obtained.

[0016] The high-precision water-soaked fingerprint recognition method and device described in the present invention utilizes a micro-humidity sensor to monitor and obtain the humidity value of the fingerprint contact area, performs image acquisition based on the humidity value, obtains a first fingerprint feature image, and utilizes the target deep learning denoising model to repair the first fingerprint feature image to obtain a second fingerprint feature image. The subcutaneous vein features of the finger are extracted to perform auxiliary verification and secondary repair on the second fingerprint feature image to obtain a final fingerprint feature image. This improves the recognition accuracy of water-soaked fingerprints and solves the problem of a high fingerprint recognition failure rate in existing fingerprint recognition modules when the finger is wet. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0018] Figure 1 This is a flow chart of a specific embodiment of a high-precision water stain fingerprint recognition method in this application; Figure 2 yes Figure 1 A flow chart of a specific embodiment of step 100; Figure 3 yes Figure 2 A schematic flow chart of a specific embodiment after step 120; Figure 4 yes Figure 3 Another specific embodiment flow chart after step 120; Figure 5 yes Figure 3 A flowchart of another specific embodiment after step 120; Figure 6 yes Figure 1A schematic flow chart of a specific embodiment of step 200; Figure 7 yes Figure 6 A schematic flow chart of a specific embodiment after step 220; Figure 8 This is a schematic diagram of the module structure of a high-precision water stain fingerprint recognition device in this application. DETAILED DESCRIPTION

[0019] The following will clearly and completely describe the technical solutions of the present invention in conjunction with the accompanying drawings. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, other embodiments obtained by ordinary technicians in this field without creative work are all within the scope of protection of the present invention.

[0020] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one skilled in the art to which this invention pertains. The terms used in this specification of the present invention are for the purpose of describing specific embodiments only and are not intended to limit the present invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0021] It should be noted that when an element is referred to as being “fixed on” or “disposed on” another element, it may be directly on the other element or indirectly on the other element. When an element is referred to as being “connected to” another element, it may be directly connected to the other element or indirectly connected to the other element.

[0022] It should be understood that the terms "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", etc., indicating the orientation or position relationship, are based on the orientation or position relationship shown in the accompanying drawings, and are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on this application.

[0023] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of the technical features being referred to. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. Throughout the description of this application, "plurality" means two or more, unless otherwise specifically defined.

[0024] Existing fingerprint recognition modules can easily fail to recognize fingerprints when there is moisture on the finger.

[0025] To solve the above problems, a high-precision water stain fingerprint recognition method and device are proposed.

[0026] In the first aspect, a high-precision water stain fingerprint recognition method is provided. Figure 1 , Figure 1 This is a flow chart of a specific embodiment of a high-precision watermark fingerprint recognition method in this application; it includes: Step 100: Obtain a micro humidity sensor, use the micro humidity sensor to monitor and obtain the humidity value of the fingerprint contact area, perform image acquisition based on the humidity value, and obtain a first fingerprint feature image.

[0027] In a preferred embodiment, Figure 2 , Figure 2 yes Figure 1 A flow chart of a specific embodiment of step 100; step 100 includes: step 110, obtaining a micro humidity sensor; step 120, embedding the micro humidity sensor on the surface of the fingerprint recognition area.

[0028] In this embodiment, a micro humidity sensor is arranged on the surface or edge of the fingerprint recognition area to monitor the humidity of the area in real time.

[0029] In a preferred embodiment, Figure 3 , Figure 3 yes Figure 2 A flow chart of a specific embodiment after step 120 in step 100; step 100 also includes step 130, obtaining a humidity value; step 140, if the humidity value is less than a specified threshold, using a visible light image acquisition mode to obtain a first fingerprint feature image.

[0030] In this embodiment, when the humidity is ≤60%, a conventional visible light image acquisition mode is adopted, and a common capacitive sensor can be used to perform fingerprint feature scanning.

[0031] In a preferred embodiment, Figure 4 , Figure 4 yes Figure 3 Schematic diagram of another specific embodiment flow chart after step 120; step 100 also includes step 150, obtaining a humidity value; step 160, if the humidity value is greater than a specified threshold, using infrared penetration imaging to capture a fingerprint image, and using an enhanced capacitive sensor to perform fingerprint feature scanning to enhance the fingerprint features on the fingerprint image to obtain a first fingerprint feature image.

[0032] In this embodiment, when the humidity is ≥60%, infrared penetrating imaging is used to capture fingerprint images, and an enhanced capacitive sensor is used to scan fingerprint features.

[0033] In a preferred embodiment, Figure 5 , Figure 5 yes Figure 3 A flow chart of another specific embodiment after step 120 in step 100 further includes: step 170, obtaining a humidity value; step 180, if the humidity value is greater than a specified threshold, prompting the user through screen display or vibration to adjust the fingerprint touch position.

[0034] Step 200: Obtain a deep learning denoising model, train the deep learning denoising model to obtain a target deep learning denoising model, use the target deep learning denoising model to repair the first fingerprint feature image to obtain a second fingerprint feature image, extract the finger subcutaneous vein features to perform auxiliary verification and secondary repair on the second fingerprint feature image to obtain a final fingerprint feature image.

[0035] In a preferred embodiment, Figure 6 , Figure 6 yes Figure 1 A flowchart of a specific embodiment of step 200 in FIG. 2 is provided; step 200 includes step 210, obtaining an input layer, multiple convolutional layers, a batch normalization layer, an activation function layer, and an output layer; step 220, constructing a deep learning denoising model using the input layer, multiple convolutional layers, a batch normalization layer, an activation function layer, and an output layer.

[0036] Input layer: accepts a noisy image as input and represents the image as a three-dimensional tensor. The dimensions are usually (height, width, channels). For color images, the number of channels is 3 (corresponding to the three RGB channels respectively), and for grayscale images, the number of channels is 1.

[0037] Convolutional layer: Consists of multiple stacked convolutional layers. Each convolutional layer contains multiple convolution kernels, which extract image features through convolution operations. The convolution kernels slide across the image, performing multiplication and addition operations with local areas of the image to generate feature maps. For example, the first convolutional layer may have 64 convolution kernels of size 3×3, with a stride of 1 and padding of 1. This allows for the extraction of preliminary image features without changing the image size. Subsequent convolutional layers gradually increase the number of feature maps to extract more complex features.

[0038] Batch Normalization layer: Located after the convolutional layer, it is used to normalize the feature maps output by the convolutional layer. It can accelerate model training, reduce the problem of vanishing or exploding gradients, and improve the model's generalization ability.

[0039] Activation function layer: The ReLU (Rectified Linear Unit) activation function is typically used. It increases the model's nonlinear expressiveness, enabling it to learn more complex functional relationships. The ReLU function is expressed as f(x)=\max(0, x). For an input value x, if x is greater than 0, the output is x; otherwise, the output is 0.

[0040] Output layer: The output of the last convolutional layer serves as the output layer. The number of convolution kernels in this layer is the same as the number of channels in the input image. Through the convolution operation, the feature map is mapped back to the same dimension as the input image, resulting in a denoised fingerprint feature image. This multi-layer convolutional structure automatically learns the mapping relationship between noisy and clean images, effectively removing noise from the image and restoring a clear image.

[0041] In a preferred embodiment, Figure 7 , Figure 7 yes Figure 6 A flowchart of a specific embodiment after step 220 in FIG. 2 is provided; step 200 further includes: step 230, collecting and obtaining fingerprint image sample data of different noise types, and obtaining a mean square error function; step 240, inputting the fingerprint image sample data into a deep learning denoising model for training, and using the mean square error function to evaluate the difference between the predicted denoised fingerprint image and the real clean fingerprint image, and correcting the parameters of the deep learning denoising model according to the evaluation result; step 250, after multiple rounds of training, obtaining a target deep learning denoising model.

[0042] In this embodiment, a large number of clean fingerprint images and corresponding noisy fingerprint images are collected. Noisy fingerprint images can be generated by adding noise of different types and intensities to the clean fingerprint images, such as Gaussian noise, salt and pepper noise, etc. The data set is divided into a training set, a validation set, and a test set. Usually, the training set is used for model training, the validation set is used to adjust the hyperparameters of the model and monitor the training process, and the test set is used to evaluate the final performance of the model. Data preprocessing: The fingerprint image is preprocessed, such as normalization, mapping the pixel values ​​to the interval [0, 1] or [-1, 1] to accelerate the convergence of the model; data enhancement can also be performed, such as random flipping, rotation, cropping, etc., to increase the diversity of the data and prevent model overfitting.

[0043] Training data is fed into the model in batches. Forward propagation calculates the model's predictions, and the loss between the predictions and the true labels is calculated using the mean squared error (MSE) loss function. Backpropagation is then used to calculate the gradient of the loss with respect to the model parameters, and the optimizer updates the model parameters based on the gradient. This process is repeated until the preset number of training rounds is reached or the loss function converges.

[0044] Step 300: Perform fingerprint matching and recognition on the final fingerprint feature image.

[0045] In an embodiment of the present application, a micro humidity sensor is used to monitor and obtain the humidity value of the fingerprint contact area, image acquisition is performed based on the humidity value, a first fingerprint feature image is obtained, and the first fingerprint feature image is repaired using a target deep learning denoising model to obtain a second fingerprint feature image. The subcutaneous vein features of the finger are extracted to perform auxiliary verification and secondary repair on the second fingerprint feature image to obtain a final fingerprint feature image, thereby improving the recognition accuracy of water-stained fingerprints and solving the problem of high fingerprint recognition failure rate of existing fingerprint recognition modules when the finger is wet.

[0046] In a second aspect, a high-precision water stain fingerprint recognition device adopts a high-precision water stain fingerprint recognition method according to the first aspect, such as Figure 8 , Figure 8 This is a schematic diagram of the module structure of a high-precision water fingerprint recognition device in the present application. It includes: an image acquisition module 401, which is used to obtain a micro-humidity sensor, monitor the humidity value of the fingerprint contact area using the micro-humidity sensor, and perform image acquisition based on the humidity value to obtain a first fingerprint feature image, wherein the micro-humidity sensor is embedded in the surface of the fingerprint recognition area; a denoising module 402, which is used to obtain a deep learning denoising model, train the deep learning denoising model to obtain a target deep learning denoising model, use the target deep learning denoising model to repair the first fingerprint feature image to obtain a second fingerprint feature image, extract the finger subcutaneous vein features to perform auxiliary verification and secondary repair on the second fingerprint feature image, and obtain a final fingerprint feature image; and an identification module 403, which is used to perform fingerprint matching and identification on the final fingerprint feature image.

[0047] Furthermore, the image acquisition module 401 includes a first image acquisition unit, a second image acquisition unit and a prompt unit; the first image acquisition unit is used to obtain the humidity value. When the humidity value is less than the specified threshold, the visible light image acquisition mode is adopted to obtain the first fingerprint feature image; the second image acquisition unit is used to obtain the humidity value. When the humidity value is greater than the specified threshold, infrared penetration imaging is adopted to acquire the fingerprint image, and an enhanced capacitive sensor is used to perform fingerprint feature scanning to enhance the fingerprint features on the fingerprint image to obtain the first fingerprint feature image; the prompt unit is used to obtain the humidity value. When the humidity value is greater than the specified threshold, the user is prompted by screen display or vibration to adjust the fingerprint touch position.

[0048] Furthermore, the denoising module 402 includes: a model construction unit, which is used to obtain an input layer, multiple convolutional layers, a batch normalization layer, an activation function layer and an output layer, and use the input layer, multiple convolutional layers, batch normalization layers, activation function layers and output layers to construct a deep learning denoising model; a training unit, which is used to collect and obtain fingerprint image sample data of different noise types, and obtain a mean square error function, input the fingerprint image sample data into the deep learning denoising model for training, and use the mean square error function to evaluate the difference between the predicted denoised fingerprint image and the real clean fingerprint image, and correct the parameters of the deep learning denoising model according to the evaluation results. After multiple rounds of training, the target deep learning denoising model is obtained.

[0049] The high-precision water-soaked fingerprint recognition method and device of the present invention utilizes a micro-humidity sensor to monitor and obtain the humidity value of the fingerprint contact area, performs image acquisition based on the humidity value, obtains a first fingerprint feature image, and utilizes a target deep learning denoising model to repair the first fingerprint feature image to obtain a second fingerprint feature image. The subcutaneous vein features of the finger are extracted to perform auxiliary verification and secondary repair on the second fingerprint feature image to obtain a final fingerprint feature image. This improves the recognition accuracy of water-soaked fingerprints and solves the problem of a high fingerprint recognition failure rate in existing fingerprint recognition modules when the finger is wet.

[0050] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A high-precision water stain fingerprint recognition method, characterized in that: include: Step 100: Obtain a micro humidity sensor, use the micro humidity sensor to monitor and obtain a humidity value of the fingerprint contact area, perform image acquisition based on the humidity value, and obtain a first fingerprint feature image; Step 200: Obtain a deep learning denoising model, train the deep learning denoising model to obtain a target deep learning denoising model, use the target deep learning denoising model to repair the first fingerprint feature image to obtain a second fingerprint feature image, extract the finger subcutaneous vein features to perform auxiliary verification and secondary repair on the second fingerprint feature image to obtain a final fingerprint feature image; Step 300: Perform fingerprint matching and recognition on the final fingerprint feature image.

2. The high-precision water stain fingerprint recognition method according to claim 1, characterized in that: The step 100 includes: Step 110: Obtain a micro humidity sensor; Step 120: embed the micro humidity sensor on the surface of the fingerprint recognition area.

3. The high-precision water stain fingerprint recognition method according to claim 2, characterized in that: The step 100 further includes: Step 130: Obtain the humidity value; Step 140: If the humidity value is less than a specified threshold, a visible light image acquisition mode is adopted to obtain the first fingerprint feature image.

4. The high-precision water stain fingerprint recognition method according to claim 2, characterized in that: The step 100 further includes: Step 150: Obtain the humidity value; Step 160: If the humidity value is greater than a specified threshold, infrared penetration imaging is used to capture a fingerprint image, and an enhanced capacitive sensor is used to perform fingerprint feature scanning to enhance the fingerprint features on the fingerprint image to obtain the first fingerprint feature image.

5. The high-precision water stain fingerprint recognition method according to claim 2, characterized in that: The step 100 further includes: Step 170: Obtain the humidity value; Step 180: If the humidity value is greater than a specified threshold, the user is prompted to adjust the fingerprint touch position through screen display or vibration.

6. The high-precision water fingerprint recognition method according to claim 1, characterized in that: The step 200 includes: Step 210: Obtain an input layer, multiple convolutional layers, a batch normalization layer, an activation function layer, and an output layer; Step 220: Build a deep learning denoising model using the input layer, multiple convolutional layers, batch normalization layer, activation function layer, and output layer.

7. The high-precision water stain fingerprint recognition method according to claim 6, characterized in that: The step 200 further includes: Step 230: Collect and obtain fingerprint image sample data of different noise types, and obtain a mean square error function; Step 240: Input the fingerprint image sample data into the deep learning denoising model for training, and use the mean square error function to evaluate the difference between the predicted denoised fingerprint image and the true clean fingerprint image, and modify the parameters of the deep learning denoising model based on the evaluation result; Step 250: After multiple rounds of training, the target deep learning denoising model is obtained.

8. A high-precision watermark fingerprint recognition device, using the high-precision watermark fingerprint recognition method according to any one of claims 1 to 7, characterized in that: include: An image acquisition module, configured to acquire a micro humidity sensor, monitor and acquire a humidity value of the fingerprint contact area using the micro humidity sensor, and acquire an image based on the humidity value to obtain a first fingerprint feature image, wherein the micro humidity sensor is embedded in the surface of the fingerprint recognition area; a denoising module, configured to obtain a deep learning denoising model, train the deep learning denoising model to obtain a target deep learning denoising model, use the target deep learning denoising model to repair the first fingerprint feature image to obtain a second fingerprint feature image, extract the finger subcutaneous vein features to perform auxiliary verification and secondary repair on the second fingerprint feature image to obtain a final fingerprint feature image; The recognition module is used to perform fingerprint matching recognition on the final fingerprint feature image.

9. The high-precision water fingerprint recognition device according to claim 8, characterized in that: The image acquisition module includes: A first image acquisition unit, a second image acquisition unit and a prompt unit; The first image acquisition unit is configured to acquire the humidity value, and when the humidity value is less than a specified threshold, adopt a visible light image acquisition mode to obtain the first fingerprint feature image; The second image acquisition unit is configured to acquire the humidity value, and when the humidity value is greater than a specified threshold, acquire a fingerprint image using infrared penetration imaging, and enhance the fingerprint features on the fingerprint image by performing fingerprint feature scanning using an enhanced capacitive sensor to obtain the first fingerprint feature image; The prompting unit is used to obtain the humidity value, and when the humidity value is greater than a specified threshold, prompt the user through screen display or vibration to adjust the fingerprint touch position.

10. The high-precision water fingerprint recognition device according to claim 8, characterized in that: The denoising module includes: A model construction unit, configured to obtain an input layer, multiple convolutional layers, a batch normalization layer, an activation function layer, and an output layer, and construct a deep learning denoising model using the input layer, multiple convolutional layers, batch normalization layers, activation function layers, and output layer; A training unit is used to collect and obtain fingerprint image sample data of different noise types and obtain a mean square error function, input the fingerprint image sample data into the deep learning denoising model for training, and use the mean square error function to evaluate the difference between the predicted denoised fingerprint image and the real clean fingerprint image, and correct the parameters of the deep learning denoising model according to the evaluation result. After multiple rounds of training, the target deep learning denoising model is obtained.

Images