Adaptive fingerprint identification method and module based on multi-modal imaging

Abstract

The invention discloses a self-adaptive fingerprint identification method and module based on multi-modal imaging, and the method comprises the steps: obtaining a multi-spectral fingerprint image sequence, triggering a plurality of light sources in a time-sharing manner, and collecting a spectrum corresponding to each light source; acquiring a stain type and a regional mask image, inputting the multispectral fingerprint image sequence into a target light convolutional neural network model, and performing stain type identification by the target light convolutional neural network model according to the multispectral fingerprint image sequence; and performing adaptive fingerprint identification, performing image enhancement according to the stain type to obtain a first enhanced image, performing local dynamic exposure compensation on the first enhanced image to obtain a second enhanced image, and transmitting the second enhanced image to a fingerprint identification unit for identification. The fingerprint recognition accuracy is improved, and the problems that the effect is poor when a traditional enhanced image software algorithm is used for recognition, and the fingerprint recognition accuracy is low due to the fact that dynamic stain type judgment is not combined are solved.

Description

Technical Field

[0001] The present invention relates to the field of fingerprint recognition technology, and in particular to an adaptive fingerprint recognition method and module based on multimodal imaging. 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] Optical fingerprint modules are easily affected by stains on the surface of the finger (such as oil stains, water stains, and dust), resulting in blurred imaging and loss of features, which in turn leads to fingerprint recognition errors.

[0004] Existing technologies primarily enhance images through software algorithms (such as contrast stretching and filtering for denoising), but this approach is limited when heavy occlusion occurs. Some solutions attempt multispectral imaging, but these methods lack dynamic occlusion detection and pose high hardware costs. Summary of the Invention

[0005] In the existing technology, when the fingerprint module is disturbed by stains on the finger surface, it uses traditional enhanced image software algorithms for recognition, but the effect is poor. When multispectral imaging is used, it is not combined with dynamic stain type judgment, and the hardware cost is relatively high.

[0006] To address the above problems, an adaptive fingerprint recognition method and module based on multimodal imaging is proposed. By time-sharing triggering of multiple light sources, a multispectral fingerprint image sequence is acquired. The target lightweight convolutional neural network model is used to identify the stain type, obtain the stain type and regional mask image, and perform image enhancement according to the stain type to obtain a first enhanced image. Local dynamic exposure compensation is performed on the first enhanced image to obtain a second enhanced image. This improves the fingerprint recognition accuracy and solves the problem of poor recognition effect of traditional enhanced image software algorithm and low fingerprint recognition accuracy caused by the lack of dynamic stain type judgment.

[0007] In a first aspect, an adaptive fingerprint recognition method based on multimodal imaging comprises: Step 100: Acquire a multispectral fingerprint image sequence, acquire an image acquisition unit, and trigger multiple light sources in a time-sharing manner. The image acquisition unit acquires the spectrum corresponding to each light source. Step 200: Obtain the stain type and the regional mask image, obtain a lightweight convolutional neural network model, train the lightweight convolutional neural network model to obtain a target lightweight convolutional neural network model, input the multispectral fingerprint image sequence into the target lightweight convolutional neural network model, and the target lightweight convolutional neural network model performs stain type recognition based on the multispectral fingerprint image sequence; Step 300, adaptive fingerprint recognition, performing image enhancement according to the stain type to obtain a first enhanced image, performing local dynamic exposure compensation on the first enhanced image to obtain a second enhanced image, and transmitting the second enhanced image to the fingerprint recognition unit for recognition.

[0008] In conjunction with the adaptive fingerprint recognition method based on multimodal imaging described in the first aspect, in a first possible implementation, step 100 includes: Step 110: Integrate a visible light source unit, a near-infrared light source unit, and a short-wave infrared light source unit into the optical fingerprint module; Step 120: Time-sharingly drive the visible light source unit, the near-infrared light source unit, and the short-wave infrared light source unit to respectively generate visible light, near-infrared light, and short-wave infrared light of specified wavelengths for irradiation; Step 130: The image acquisition unit acquires the corresponding fingerprint image spectrum to obtain the multispectral fingerprint image sequence.

[0009] In conjunction with the adaptive fingerprint recognition method based on multimodal imaging described in the first aspect, in a second possible implementation, step 200 includes: Step 210: Obtain an input layer, multiple convolutional layers, multiple pooling layers, and an output layer; Step 220: Utilize the input layer, multiple convolutional layers, multiple pooling layers, and output layer to construct the lightweight convolutional neural network model.

[0010] In combination with the second possible implementation of the first aspect, in a third possible implementation, step 200 includes: Step 230: Collect image data samples of multiple stain types and pre-process the image data samples; Step 240: Input the preprocessed image data samples into the lightweight convolutional neural network model for training to obtain a target lightweight convolutional neural network model.

[0011] In conjunction with the adaptive fingerprint recognition method based on multimodal imaging described in the first aspect, in a fourth possible implementation, step 300 includes: Step 310: Adjust the outputs of the multiple light sources according to the type of stain to enhance fingerprint features in the fingerprint image to obtain the first enhanced image; Step 320: Adjust exposure parameters for each region based on the brightness distribution of the stain area in the first enhanced image, perform exposure compensation on the fingerprint features of the stain area, and obtain the second enhanced image.

[0012] In combination with the fourth possible implementation of the first aspect, in a fifth possible implementation, step 310 includes: Step 311: If the image is of oil stain type, increase the channel weight of the near-infrared light source to suppress the reflection noise of the visible light source; Step 312: If the fingerprint is of the water stain type, increase the channel weight of the short-wave infrared light source to enhance the edge features of the fingerprint image, turn on the visible light source normally to obtain the detail features of the fingerprint image, and combine the edge features with the detail features; Step 313: If the type is dust, the visible light source, near-infrared light source and short-wave infrared light source are normally turned on to obtain a three-spectrum fingerprint image, and the fingerprint features of the three-spectrum fingerprint image are averaged and fused to reduce the influence of local occlusion.

[0013] In combination with the fourth possible implementation of the first aspect, in a sixth possible implementation, step 320 includes: Step 321: If the stain area of ​​the first enhanced image is a high-reflection area, reduce the exposure gain to avoid overexposure; Step 322: If the stain area of ​​the first enhanced image is a low-reflection area, extend the exposure time to enhance texture features.

[0014] In a second aspect, an adaptive fingerprint recognition module based on multimodal imaging is provided, which uses the adaptive fingerprint recognition method based on multimodal imaging described in the first aspect for recognition, comprising: The first acquisition module is used to acquire a multispectral fingerprint image sequence, trigger multiple light sources in time-sharing, and use the image acquisition unit to collect the spectrum corresponding to each light source; a second acquisition module, configured to acquire the stain type and the regional mask image, train the lightweight convolutional neural network model to obtain a target lightweight convolutional neural network model, input the multispectral fingerprint image sequence into the target lightweight convolutional neural network model, and the target lightweight convolutional neural network model performs stain type recognition based on the multispectral fingerprint image sequence; The adaptive recognition module is used to perform adaptive fingerprint recognition, perform image enhancement according to the stain type to obtain a first enhanced image, perform local dynamic exposure compensation on the first enhanced image to obtain a second enhanced image, and transmit the second enhanced image to the fingerprint recognition unit for recognition.

[0015] In conjunction with the adaptive fingerprint recognition module based on multimodal imaging described in the second aspect, in a first possible implementation, the adaptive recognition module includes: a first image enhancement unit, configured to adjust the outputs of the plurality of light sources according to the type of stain, so as to enhance fingerprint features in the fingerprint image and obtain the first enhanced image; The second image enhancement unit is used to adjust the exposure parameters of the brightness distribution of the stain area of ​​the first enhanced image by region, perform exposure compensation on the fingerprint features of the stain area, and obtain the second enhanced image.

[0016] In combination with the first possible implementation of the second aspect, in a second possible implementation, the adaptive recognition module further includes: An adaptive recognition unit is configured to perform fingerprint recognition on the second enhanced image.

[0017] The adaptive fingerprint recognition method and module based on multimodal imaging described in the present invention are implemented to obtain a multispectral fingerprint image sequence by time-sharingly triggering multiple light sources, using a target lightweight convolutional neural network model to distinguish the stain type, obtain the stain type and regional mask image, and perform image enhancement according to the stain type to obtain a first enhanced image. Local dynamic exposure compensation is performed on the first enhanced image to obtain a second enhanced image, thereby improving the fingerprint recognition accuracy and solving the problem of poor recognition effect of traditional enhanced image software algorithms and low fingerprint recognition accuracy caused by the lack of dynamic stain type judgment. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] 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.

[0019] Figure 1 This is a flowchart of a specific embodiment of an adaptive fingerprint recognition method based on multimodal imaging in this application; Figure 2 yes Figure 1 A schematic flow chart of a specific embodiment of step 100; Figure 3 yes Figure 1 A schematic flow chart of a specific embodiment of step 200; Figure 4 yes Figure 3 A schematic flow chart of a specific embodiment after step 220; Figure 5 yes Figure 1A flow chart of a specific embodiment of step 300; Figure 6 yes Figure 5 A schematic flow chart of a specific embodiment of step 310; Figure 7 yes Figure 5 A schematic flow chart of a specific embodiment of step 320; Figure 8 This is a schematic diagram of the module structure of an adaptive fingerprint recognition device based on multimodal imaging in this application. DETAILED DESCRIPTION

[0020] 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.

[0021] 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.

[0022] 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.

[0023] 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.

[0024] 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.

[0025] In the existing technology, when the fingerprint module is disturbed by stains on the finger surface, it uses traditional enhanced image software algorithms for recognition, but the effect is poor. When multispectral imaging is used, it is not combined with dynamic stain type judgment, and the hardware cost is relatively high.

[0026] To address the above problems, an adaptive fingerprint recognition method and module based on multimodal imaging are proposed.

[0027] First, an adaptive fingerprint recognition method based on multimodal imaging, please refer to Figure 1 , Figure 1 This is a flow chart of a specific embodiment of an adaptive fingerprint recognition method based on multimodal imaging in this application; it includes: Step 100: Acquire a multispectral fingerprint image sequence, acquire an image acquisition unit, and trigger multiple light sources in a time-sharing manner. The image acquisition unit acquires the spectrum corresponding to each light source.

[0028] In a preferred embodiment, please refer to Figure 2 , Figure 2 yes Figure 1 A flow chart of a specific embodiment of step 100 in FIG. 1 is provided; step 100 includes: step 110, integrating a visible light source unit, a near-infrared light source unit, and a short-wave infrared light source unit in an optical fingerprint module; step 120, time-sharing driving the visible light source unit, the near-infrared light source unit, and the short-wave infrared light source unit to respectively generate visible light, near-infrared light, and short-wave infrared light of specified wavelengths for irradiation; step 130, the image acquisition unit acquires the corresponding fingerprint image spectrum to obtain a multispectral fingerprint image sequence.

[0029] In this embodiment, the fingerprint recognition module integrates a visible light source unit, a near-infrared light source unit, and a short-wave infrared light source unit. The wavelengths of visible light are 520-610nm, near-infrared light is 850nm, and short-wave infrared light is 1300nm. These three light sources are triggered in a time-sharing manner. Different wavelengths have different stain penetration properties (for example, near-infrared light penetrates oil stains more effectively), generating a multispectral fingerprint image sequence.

[0030] Step 200: Obtain the stain type and regional mask image, obtain a lightweight convolutional neural network model, train the lightweight convolutional neural network model to obtain a target lightweight convolutional neural network model, input the multispectral fingerprint image sequence into the target lightweight convolutional neural network model, and the target lightweight convolutional neural network model identifies the stain type based on the multispectral fingerprint image sequence.

[0031] In a preferred embodiment, please refer to Figure 3 , Figure 3 yes Figure 1 A flow chart of a specific embodiment of step 200; step 200 includes: step 210, obtaining an input layer, multiple convolutional layers, multiple pooling layers and an output layer; step 220, using the input layer, multiple convolutional layers, multiple pooling layers and the output layer to construct a lightweight convolutional neural network model.

[0032] In this embodiment, the lightweight convolutional neural network model can be: 1. Input layer; 2. Convolutional layer 1: convolution kernel size 3x3, number of convolution kernels is 32, stride is 1, padding is same, activation function is ReLU; 3. Pooling layer 1: The pooling method is maximum pooling, the pooling kernel size is 2×2, and the stride is 2. The output of convolution layer 1 is downsampled to reduce the data dimension while retaining the main features; 4. Convolutional layer 2: convolution kernel size 3x3, number of convolution kernels is 64, stride is 1, padding is same, activation function is ReLU; 5. Pooling layer 2: The pooling method is maximum pooling, the pooling kernel size is 2×2, and the stride is 2. The output of convolutional layer 2 is downsampled to reduce the data dimension while retaining the main features; 6. Convolutional layer 3: convolution kernel size 3x3, number of convolution kernels is 128, stride is 1, padding is same, activation function is ReLU; 7. Pooling layer 3: The pooling method is maximum pooling, the pooling kernel size is 2×2, and the stride is 2. The output of convolution layer 3 is downsampled to reduce the data dimension while retaining the main features; 8. Fully connected layer 1: Flattens the feature map output by pooling layer 3 into a one-dimensional vector and inputs it into the fully connected layer for further feature combination and abstraction. The activation function is ReLU. 9. Fully connected layer 2: The number of neurons corresponds to the number of stain types (n+1=4). The activation function is Softmax, which is used to output the probability distribution of each category to determine the probability that the input spectral fingerprint image belongs to each stain type or clean type.

[0033] In a preferred embodiment, please refer to Figure 4 , Figure 4 yes Figure 3 A flowchart of a specific embodiment after step 220 in FIG. 200 includes: step 230, collecting image data samples of multiple stain types and preprocessing the image data samples; step 240, inputting the preprocessed image data samples into a lightweight convolutional neural network model for training to obtain a target lightweight convolutional neural network model.

[0034] In this embodiment, a large amount of fingerprint image data of different stain types (oil stains, water stains, and dust) as well as clean types are collected to ensure that the image data samples are representative and diverse. The stain type of each image data is marked, and image enhancement and generalization are performed through rotation, scaling, evaluation, etc.

[0035] In this embodiment, a cross entropy loss function can be selected to measure the difference between the model prediction result and the true label.

[0036] Step 300, adaptive fingerprint recognition, performing image enhancement according to the stain type to obtain a first enhanced image, performing local dynamic exposure compensation on the first enhanced image to obtain a second enhanced image, and transmitting the second enhanced image to the fingerprint recognition unit for recognition.

[0037] In a preferred embodiment, please refer to Figure 5 , Figure 5 yes Figure 1 A flow chart of a specific embodiment of step 300 is shown in FIG. 3 ; step 300 includes: Step 310: Adjust the outputs of the multiple light sources according to the stain type to enhance the fingerprint features in the fingerprint image to obtain a first enhanced image.

[0038] In a preferred embodiment, please refer to Figure 6 , Figure 6 yes Figure 5 A flow chart of a specific embodiment of step 310 in FIG. 310 ; step 310 includes: step 311, if it is an oil stain type, then increasing the channel weight of the near-infrared light source to suppress the reflection noise of the visible light source; step 312, if it is a water stain type, then increasing the channel weight of the short-wave infrared light source to enhance the edge features of the fingerprint image, turning on the visible light source normally to obtain the detail features of the fingerprint image, and combining the edge features with the detail features; step 313, if it is a dust type, then turning on the visible light source, the near-infrared light source and the short-wave infrared light source normally to obtain a three-spectrum fingerprint image, and performing mean fusion on the fingerprint features of the three-spectrum fingerprint image to reduce the impact of local occlusion.

[0039] Step 320: Adjust exposure parameters for each region based on the brightness distribution of the stain area in the first enhanced image, perform exposure compensation on the fingerprint features of the stain area, and obtain a second enhanced image.

[0040] In a preferred embodiment, please refer to Figure 7 , Figure 7 yes Figure 5 FIG3 is a flow chart of a specific embodiment of step 320 in FIG3; step 320 includes: step 321, if the stain area of ​​the first enhanced image is a high-reflection area, reducing the exposure gain to avoid overexposure; step 322, if the stain area of ​​the first enhanced image is a low-reflection area, extending the exposure time to enhance the texture features.

[0041] In this embodiment, a multispectral fingerprint image sequence is obtained by time-sharing triggering multiple light sources, and the target lightweight convolutional neural network model is used to distinguish the stain type, obtain the stain type and regional mask image, and perform image enhancement according to the stain type to obtain a first enhanced image. Local dynamic exposure compensation is performed on the first enhanced image to obtain a second enhanced image, thereby improving the fingerprint recognition accuracy and solving the problem of poor recognition effect of traditional enhanced image software algorithms and low fingerprint recognition accuracy due to the lack of dynamic stain type judgment.

[0042] In the second aspect, an adaptive fingerprint recognition module based on multimodal imaging is used to perform recognition using the adaptive fingerprint recognition method based on multimodal imaging of the first aspect. Please refer to Figure 8 , Figure 8 This is a schematic diagram of the module structure of an adaptive fingerprint recognition device based on multimodal imaging in the present application. It includes: a first acquisition module 401, which is used to acquire a multispectral fingerprint image sequence, time-share trigger multiple light sources, and use an image acquisition unit to acquire the spectrum corresponding to each light source; a second acquisition module 402, which is used to acquire the stain type and regional mask image, train a lightweight convolutional neural network model to obtain a target lightweight convolutional neural network model, input the multispectral fingerprint image sequence into the target lightweight convolutional neural network model, and the target lightweight convolutional neural network model identifies the stain type based on the multispectral fingerprint image sequence; and an adaptive recognition module 403, which is used to perform adaptive fingerprint recognition, perform image enhancement based on the stain type to obtain a first enhanced image, perform local dynamic exposure compensation on the first enhanced image to obtain a second enhanced image, and transmit the second enhanced image to the fingerprint recognition unit for recognition.

[0043] Furthermore, the adaptive recognition module 403 includes: a first image enhancement unit, which is used to adjust the output of multiple light sources according to the type of stain to enhance the fingerprint features in the fingerprint image and obtain a first enhanced image; a second image enhancement unit, which is used to adjust the exposure parameters of the brightness distribution of the stain area of ​​the first enhanced image by region, and perform exposure compensation on the fingerprint features of the stain area to obtain a second enhanced image.

[0044] Furthermore, the adaptive recognition module 403 further includes an adaptive recognition unit configured to perform fingerprint recognition on the second enhanced image.

[0045] The present invention implements an adaptive fingerprint recognition method and module based on multimodal imaging, which acquires a multispectral fingerprint image sequence by time-sharingly triggering multiple light sources, uses a target lightweight convolutional neural network model to discriminate the stain type, obtains the stain type and regional mask image, and performs image enhancement according to the stain type to obtain a first enhanced image. Local dynamic exposure compensation is performed on the first enhanced image to obtain a second enhanced image, thereby improving the fingerprint recognition accuracy and solving the problem of poor recognition effect of traditional enhanced image software algorithms and low fingerprint recognition accuracy caused by the lack of dynamic stain type judgment.

[0046] 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. An adaptive fingerprint recognition method based on multimodal imaging, characterized in that: include: Step 100: Obtain a multispectral fingerprint image sequence: Acquire an image acquisition unit, and trigger multiple light sources in a time-sharing manner, and use the image acquisition unit to acquire the spectrum corresponding to each light source; Step 200: Obtain stain type and region mask image: Obtaining a lightweight convolutional neural network model, training the lightweight convolutional neural network model to obtain a target lightweight convolutional neural network model, inputting the multispectral fingerprint image sequence into the target lightweight convolutional neural network model, and the target lightweight convolutional neural network model performing stain type recognition based on the multispectral fingerprint image sequence; Step 300: Adaptive fingerprint recognition: Image enhancement is performed according to the stain type to obtain a first enhanced image, local dynamic exposure compensation is performed on the first enhanced image to obtain a second enhanced image, and the second enhanced image is transmitted to the fingerprint recognition unit for recognition.

2. The adaptive fingerprint recognition method based on multimodal imaging according to claim 1, characterized in that: The step 100 includes: Step 110: Integrate a visible light source unit, a near-infrared light source unit, and a short-wave infrared light source unit into the optical fingerprint module; Step 120: Time-sharingly drive the visible light source unit, the near-infrared light source unit, and the short-wave infrared light source unit to respectively generate visible light, near-infrared light, and short-wave infrared light of specified wavelengths for irradiation; Step 130: Utilize the image acquisition unit to acquire the corresponding fingerprint image spectrum to obtain the multispectral fingerprint image sequence.

3. The adaptive fingerprint recognition method based on multimodal imaging according to claim 1, characterized in that: The step 200 includes: Step 210: Obtain an input layer, multiple convolutional layers, multiple pooling layers, and an output layer; Step 220: Utilize the input layer, multiple convolutional layers, multiple pooling layers, and output layer to construct the lightweight convolutional neural network model.

4. The adaptive fingerprint recognition method based on multimodal imaging according to claim 3, characterized in that: The step 200 includes: Step 230: Collect image data samples of multiple stain types and pre-process the image data samples; Step 240: Input the preprocessed image data samples into the lightweight convolutional neural network model for training to obtain a target lightweight convolutional neural network model.

5. The adaptive fingerprint recognition method based on multimodal imaging according to claim 1, characterized in that: The step 300 includes: Step 310: Adjust the outputs of the multiple light sources according to the type of stain to enhance fingerprint features in the fingerprint image to obtain the first enhanced image; Step 320: Adjust exposure parameters for each region based on the brightness distribution of the stain area in the first enhanced image, perform exposure compensation on the fingerprint features of the stain area, and obtain the second enhanced image.

6. The adaptive fingerprint recognition method based on multimodal imaging according to claim 5, characterized in that: The step 310 includes: Step 311: If the image is of oil stain type, increase the channel weight of the near-infrared light source to suppress the reflection noise of the visible light source; Step 312: If the fingerprint is of the water stain type, increase the channel weight of the short-wave infrared light source to enhance the edge features of the fingerprint image, turn on the visible light source normally to obtain the detail features of the fingerprint image, and combine the edge features with the detail features; Step 313: If the type is dust, the visible light source, near-infrared light source and short-wave infrared light source are normally turned on to obtain a three-spectrum fingerprint image, and the fingerprint features of the three-spectrum fingerprint image are averaged and fused to reduce the influence of local occlusion.

7. The adaptive fingerprint recognition method based on multimodal imaging according to claim 5, characterized in that: The step 320 includes: Step 321: If the stain area of ​​the first enhanced image is a high-reflection area, reduce the exposure gain to avoid overexposure; Step 322: If the stain area of ​​the first enhanced image is a low-reflection area, extend the exposure time to enhance texture features.

8. An adaptive fingerprint recognition module based on multimodal imaging, which adopts the adaptive fingerprint recognition method based on multimodal imaging according to any one of claims 1 to 7 for recognition, characterized in that: include: The first acquisition module is used to acquire a multispectral fingerprint image sequence, trigger multiple light sources in time-sharing, and use the image acquisition unit to collect the spectrum corresponding to each light source; a second acquisition module, configured to acquire the stain type and the regional mask image, train the lightweight convolutional neural network model to obtain a target lightweight convolutional neural network model, input the multispectral fingerprint image sequence into the target lightweight convolutional neural network model, and the target lightweight convolutional neural network model performs stain type recognition based on the multispectral fingerprint image sequence; The adaptive recognition module is used to perform adaptive fingerprint recognition, perform image enhancement according to the stain type to obtain a first enhanced image, perform local dynamic exposure compensation on the first enhanced image to obtain a second enhanced image, and transmit the second enhanced image to the fingerprint recognition unit for recognition.

9. The adaptive fingerprint recognition module based on multimodal imaging according to claim 8, characterized in that: The adaptive recognition module includes: a first image enhancement unit, configured to adjust the outputs of the plurality of light sources according to the type of stain, so as to enhance fingerprint features in the fingerprint image and obtain the first enhanced image; The second image enhancement unit is used to adjust the exposure parameters of the brightness distribution of the stain area of ​​the first enhanced image by region, perform exposure compensation on the fingerprint features of the stain area, and obtain the second enhanced image.

10. The adaptive fingerprint recognition module based on multimodal imaging according to claim 9, characterized in that: The adaptive recognition module also includes: An adaptive recognition unit is configured to perform fingerprint recognition on the second enhanced image.

Images