Entrance guard verification fingerprint identification method based on biological identification technology

Abstract

The invention discloses an access control verification fingerprint identification method based on a biological identification technology. The method comprises the following steps: acquiring a user fingerprint image based on an access control system; calculating directions of minutiae based on the fingerprint area statistical information; based on the directions of the minutiae, adding descriptors with statistical significance for the minutiae; extracting triangular vertex directions of the minutiae; the fingerprint identification method has the advantages that the accuracy and the anti-interference capability of identification are improved by accurately extracting the features of the minutiae, statistically analyzing the gradient direction and acquiring the distribution of the gradient direction. The method shows good adaptability in a complex environment, can quickly and reliably carry out identity verification, and ensures the safety and high efficiency of an access control system.

Description

Technical Field

[0001] The invention relates to fingerprint verification, and in particular to a fingerprint recognition method for access control verification based on biometric recognition technology. Background Art

[0002] With the continuous development of information technology and biometric technology, traditional access control verification methods are gradually unable to meet the increasingly stringent security needs. Verification methods such as passwords and cards are easily stolen or lost. Therefore, access control systems based on biometric technology have become a safer and more efficient solution. As an important form of biometric technology, fingerprint recognition technology is widely used in identity authentication in various secure places due to its uniqueness, stability and ease of use. Fingerprint recognition uses the unique pattern characteristics of fingerprints to confirm identity. Everyone's fingerprint is unique and difficult to imitate, which makes fingerprint recognition a reliable identity authentication method. Compared with traditional access control verification methods, fingerprint recognition technology not only improves security, but also simplifies the verification process, avoiding the problems of forgetting, losing or misappropriation that are easy to occur in traditional methods.

[0003] The fingerprint recognition methods for access control verification currently on the market mainly include optical fingerprint recognition and capacitive fingerprint recognition. Optical fingerprint recognition uses the principle of light reflection to capture fingerprint images through a camera, while capacitive fingerprint recognition recognizes fingerprints by sensing the change in capacitance between the fingerprint and the sensor. However, the traditional expression of detail points has problems with insufficient stability and specificity, resulting in low accuracy in the fingerprint registration and matching process. Detail points (such as endpoints, bifurcation points, etc.) are key features in fingerprint images, and they play a vital role in fingerprint recognition. However, due to factors such as environmental changes, image noise, and fingerprint damage, the detail point expression method relied on by traditional methods may lack sufficient robustness, resulting in errors in the extraction and matching of detail points in complex situations. Summary of the invention

[0004] In order to improve the existing access control verification fingerprint recognition method, a access control verification fingerprint recognition method based on biometric recognition technology is provided. This method obtains the fingerprint image and extracts the direction and gradient distribution of the detail points to generate a descriptor with statistical significance to achieve accurate identity authentication. This method improves the recognition accuracy and anti-interference ability by performing four-quadrant analysis and interpolation operations on the detail point area, ensuring the efficient and stable operation of the system in complex environments.

[0005] In order to achieve the above purpose, the technical solution adopted by the present invention is: A fingerprint recognition method for access control verification based on biometrics technology, comprising: Obtain user fingerprint images based on the access control system; Calculate the direction of the minutiae points based on the statistical information of the fingerprint region; Based on the direction of detail points, add statistically significant descriptors to detail points; Extract the triangular vertex directions of detail points; Fingerprint matching is performed based on the extracted detail points to obtain fingerprint recognition results.

[0006] Preferably, the direction of calculating the minutiae points based on the fingerprint region statistical information specifically includes: Select the statistical area of ​​a certain detail point; Calculate the gradient direction of all detail points in the statistical fingerprint area; Based on the statistical gradient directions of all detail points in the fingerprint area, the gradient direction distribution of all detail points in the fingerprint area is statistically analyzed; Perform interpolation operation based on the gradient direction distribution to obtain the direction distribution function; Get the maximum value of the direction distribution function, get the direction of the detail point, and repeat the above steps until all detail points of the input fingerprint image are traversed.

[0007] Preferably, the step of calculating the gradient directions of all detail points in the statistical fingerprint region specifically includes: The Sobel operator formula for horizontal gradient is: , The Sobel operator formula for the vertical gradient is: , The horizontal gradient and vertical gradient of detail points are calculated based on the Sobel operator. The formula is: , , in, is the position of the pixel value, , are the weights of the horizontal and vertical Sobel operators, , are the horizontal and vertical gradients of detail points; Based on the horizontal and vertical gradients of the acquired detail points, the gradient direction of the detail points is calculated using the formula: , Based on this formula, the gradient direction of each detail point is calculated in radians. The gradient direction value is in the range [0, 2 ) within the range of .

[0008] Preferably, the distribution of the gradient directions of all the minutiae points in the fingerprint region based on the statistics of the gradient directions of all the minutiae points in the fingerprint region specifically includes: Based on the statistical gradient directions of all detail points in the fingerprint area, the intervals are divided and the gradient direction distribution is obtained through the direction histogram. The formula is: , in, is the gradient direction of the detail point, is an indicator function, indicating whether the gradient direction of the detail point falls within the current direction range; Based on the obtained directional histogram, normalization is performed to ensure that the sum of the directional histogram is 1. The formula is: , Based on the normalized direction histogram, the gradient direction distribution of all detail points is obtained.

[0009] Preferably, interpolation operation is performed based on the gradient direction distribution to obtain the direction distribution function, which specifically includes: Based on the acquired direction histogram, the direction distribution is smoothed by linear interpolation. Assuming that in the interval [ ) is a linear function, then the directional distribution function formula after interpolation is: , Based on the periodicity of the gradient direction of the detail points, the interpolation function is periodically extended so that exist and The values ​​at are the same.

[0010] Preferably, the adding of statistically significant descriptors to detail points specifically includes: Select the descriptor extraction area of ​​a certain detail point; Determine the four quadrants of the region based on the descriptor extraction region and the direction of the detail point; Calculate the gradient direction of all detail points in the four quadrants of the region and perform interpolation operations; The gradient distribution of the four quadrants of the region after the interpolation operation is statistically calculated to form the descriptor of the detail point, and the above steps are repeated until all detail points of the input fingerprint image are traversed.

[0011] Preferably, the extracting region based on the descriptor and determining the four quadrants of the region based on the direction of the detail point specifically includes: Based on the circular extraction area, the direction of the detail point is taken as the positive direction of the horizontal axis, the detail point is taken as the origin, and then the circular area is divided into 4 quadrants according to the plane rectangular coordinate system. The rotation equation is: , Among them, [x, y] is the original coordinate, [x', y'] is the rotated coordinate, and θ is the direction of the detail point.

[0012] Preferably, the triangular vertex directions of extracting detail points specifically include: Based on the gradient direction of the acquired detail points, a triangle model is constructed, and the angles of each branch line segment are calculated to determine the main direction. The formula is: , in, are the directions of the three branches, and their average is taken to represent the main direction of the triangle vertices.

[0013] Preferably, performing fingerprint matching based on the extracted minutiae to obtain the fingerprint recognition result specifically includes: By calculating the distance and direction similarity between detail points, the similarity of detail points is obtained. The formula is: , in, , For two details, is the Euclidean distance between two minutiae points, is the directional similarity of two detail points; When the similarity of the minutiae points meets a preset first threshold, adjusting the fingerprint pattern; The similarity of overlapping regions is calculated based on the normalized cross-correlation coefficient, and the formula is: , in, is the overlap region similarity, is the overlapping area, For two fingerprint images at position The pixel value of When the similarity of the overlapping areas meets a preset second threshold, it is determined that the match is successful.

[0014] Compared with the prior art, the advantages of the present invention are: By using minutiae extraction and gradient direction statistical analysis, the unique features in the fingerprint image can be accurately captured, thus improving the accuracy of fingerprint recognition. By accurately calculating the direction of each minutiae point and combining the gradient distribution in the region, the method effectively reduces the interference caused by noise and blur, and enhances the adaptability and robustness of fingerprint recognition in different environments. Secondly, through multi-dimensional analysis of minutiae descriptors, this method not only ensures the high recognition of the fingerprint image, but also improves the anti-interference ability of the recognition process. Even when the fingerprint is damaged or of poor quality, the system can still maintain a high recognition accuracy. In addition, when the system performs fingerprint matching, it quickly and accurately completes the verification operation through fine processing of the fingerprint image, thus ensuring the security and efficiency of the access control system. Finally, the anti-interference ability of this method makes it more reliable in practical applications and can adapt to a variety of complex environments, such as when the fingerprint is wet, stained or slightly damaged. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 A schematic diagram of fingerprint verification of the method proposed in the present invention; Figure 2 A schematic diagram of detail point direction calculation of the method proposed in the present invention; Figure 3 A schematic diagram of the directional gradient of detail points of the method proposed in the present invention; Figure 4 This is a schematic diagram of the gradient direction distribution of the method proposed in the present invention; Figure 5 A schematic diagram of the direction distribution function of the method proposed in the present invention; Figure 6 Add a schematic diagram for the descriptor of the method proposed in the present invention; Figure 7 This is a schematic diagram of the fingerprint recognition result of the method proposed in the present invention. DETAILED DESCRIPTION

[0016] The following description is used to disclose the present invention so that those skilled in the art can implement the present invention. The preferred embodiments described below are only examples, and those skilled in the art may think of other obvious variations.

[0017] A fingerprint recognition method for access control verification based on biometrics technology, comprising: Step 1: Obtain user fingerprint image based on access control system; Step 2: Calculate the direction of the minutiae points based on the statistical information of the fingerprint area; Step 3: Based on the direction of the detail points, add statistically significant descriptors to the detail points; Step 4: Extract the triangular vertex directions of detail points; Step 5: Perform fingerprint matching based on the extracted detail points to obtain fingerprint recognition results.

[0018] See also Figure 2 As shown, the direction of calculating the minutiae points based on the fingerprint area statistical information specifically includes: Select the statistical area of ​​a certain detail point; Calculate the gradient direction of all detail points in the statistical fingerprint area; Based on the statistical gradient directions of all detail points in the fingerprint area, the gradient direction distribution of all detail points in the fingerprint area is statistically analyzed; Perform interpolation operation based on the gradient direction distribution to obtain the direction distribution function; Get the maximum value of the direction distribution function, get the direction of the detail point, and repeat the above steps until all detail points of the input fingerprint image are traversed.

[0019] See also Figure 3 As shown, the calculation of the gradient direction of all detail points in the statistical fingerprint area specifically includes: The Sobel operator formula for horizontal gradient is: , The Sobel operator formula for the vertical gradient is: , The horizontal gradient and vertical gradient of detail points are calculated based on the Sobel operator. The formula is: , , in, is the position of the pixel value, , are the weights of the horizontal and vertical Sobel operators, , are the horizontal and vertical gradients of detail points; Based on the horizontal and vertical gradients of the acquired detail points, the gradient direction of the detail points is calculated using the formula: , Based on this formula, the gradient direction of each detail point is calculated in radians. The gradient direction value is in the range [0, 2 ) within the range of .

[0020] Specifically, in order to enhance the influence of important areas in the fingerprint image and reduce the interference of irrelevant areas or noise on the recognition results, weighted gradients can be used to give higher weights to more recognizable areas such as edges and ridges in the fingerprint image, thereby improving matching accuracy. At the same time, weighted gradients can effectively suppress the interference of noise or unimportant information in the fingerprint image, especially when the fingerprint image quality is poor or there are problems such as contamination or blur. Through weighted processing, the system can more accurately identify the direction and position of detail points and improve the robustness of fingerprint recognition.

[0021] See also Figure 4 As shown, based on the gradient directions of all detail points in the statistical fingerprint area, the distribution of the gradient directions of all detail points in the statistical fingerprint area specifically includes: Based on the statistical gradient directions of all detail points in the fingerprint area, the intervals are divided and the gradient direction distribution is obtained through the direction histogram. The formula is: , in, is the gradient direction of the detail point, is an indicator function, indicating whether the gradient direction of the detail point falls within the current direction range; Based on the obtained directional histogram, normalization is performed to ensure that the sum of the directional histogram is 1. The formula is: , Based on the normalized direction histogram, the gradient direction distribution of all detail points is obtained.

[0022] It is understandable that the fingerprint image may be rotated or distorted due to factors such as the acquisition angle and compression, which will affect the gradient direction distribution statistics of the detail points. Especially in the case of inconsistent rotation, the gradient direction statistics will be biased. Therefore, before calculating the gradient direction distribution, the image rotation correction is performed to ensure the positivity of the fingerprint image and reduce the impact of the rotation. A rotation-invariant algorithm can also be used to extract detail points (such as technology based on direction fields) to avoid the interference of rotation on the gradient direction distribution.

[0023] See also Figure 5 As shown, interpolation operation is performed based on the gradient direction distribution, and the direction distribution function obtained specifically includes: Based on the acquired direction histogram, the direction distribution is smoothed by linear interpolation. Assuming that in the interval [ ) is a linear function, then the directional distribution function formula after interpolation is: , Based on the periodicity of the gradient direction of the detail points, the interpolation function is periodically extended so that exist and The values ​​at are the same.

[0024] See also Figure 6 As shown in Figure 2, adding statistically significant descriptors to detail points specifically includes: Select the descriptor extraction area of ​​a certain detail point; Determine the four quadrants of the region based on the descriptor extraction region and the direction of the detail point; Calculate the gradient direction of all detail points in the four quadrants of the region and perform interpolation operations; The gradient distribution of the four quadrants of the region after the interpolation operation is statistically calculated to form the descriptor of the detail point, and the above steps are repeated until all detail points of the input fingerprint image are traversed.

[0025] The four quadrants of the region are determined based on the descriptor extraction region and the direction of the detail point, specifically including: Based on the circular extraction area, the direction of the detail point is taken as the positive direction of the horizontal axis, the detail point is taken as the origin, and then the circular area is divided into 4 quadrants according to the plane rectangular coordinate system. The rotation equation is: , Among them, [x, y] is the original coordinate, [x', y'] is the rotated coordinate, and θ is the direction of the detail point.

[0026] The triangular vertex directions for extracting detail points specifically include: Based on the gradient direction of the acquired detail points, a triangle model is constructed, and the angles of each branch line segment are calculated to determine the main direction. The formula is: , in, are the directions of the three branches, and their average is taken to represent the main direction of the triangle vertices.

[0027] See also Figure 7 As shown, fingerprint matching is performed based on the extracted minutiae points, and the fingerprint recognition results are obtained specifically including: By calculating the distance and direction similarity between detail points, the similarity of detail points is obtained. The formula is: , in, , For two details, is the Euclidean distance between two minutiae points, is the directional similarity of two detail points; When the similarity of the minutiae points meets a preset first threshold, adjusting the fingerprint pattern; The similarity of overlapping regions is calculated based on the normalized cross-correlation coefficient, and the formula is: , in, is the overlap region similarity, is the overlapping area, For two fingerprint images at position The pixel value of When the similarity of the overlapping areas meets a preset second threshold, it is determined that the match is successful.

[0028] It is understandable that due to the limitations of the method for calculating the similarity of minutiae points, mismatches or false matches may occur, especially between fingerprint pairs with high similarity, where the possibility of mismatches increases. In addition to minutiae point matching, other features, such as texture features, directional field features, etc., can also be used for auxiliary matching to increase the accuracy of matching. At the same time, after matching, post-processing algorithms such as conditional random fields are used to further optimize the matching results and reduce false matches.

[0029] It should be noted that the sequence of the above embodiments of the present invention is only for description and does not represent the advantages and disadvantages of the embodiments. The above is a description of a specific embodiment of this specification. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0030] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

[0031] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A fingerprint recognition method for access control verification based on biometric technology, characterized in that: include: Obtain user fingerprint images based on the access control system; Calculate the direction of the minutiae points based on the statistical information of the fingerprint region; Based on the direction of detail points, add statistically significant descriptors to detail points; Extract the triangular vertex directions of detail points; Fingerprint matching is performed based on the extracted detail points to obtain fingerprint recognition results.

2. According to the biometric access control fingerprint recognition method of claim 1, it is characterized in that: The direction of calculating the minutiae points based on the fingerprint region statistical information specifically includes: Select the statistical area of ​​a certain detail point; Calculate the gradient direction of all detail points in the statistical fingerprint area; Based on the statistical gradient directions of all detail points in the fingerprint area, the distribution of gradient directions of all detail points in the fingerprint area is statistically analyzed; Perform interpolation operation based on the gradient direction distribution to obtain the direction distribution function; Get the maximum value of the direction distribution function, get the direction of the detail point, and repeat the above steps until all detail points of the input fingerprint image are traversed.

3. According to the biometric access control fingerprint recognition method of claim 2, it is characterized in that: The calculation of the gradient directions of all minutiae points in the statistical fingerprint area specifically includes: The Sobel operator formula for horizontal gradient is: , The Sobel operator formula for the vertical gradient is: ， The horizontal gradient and vertical gradient of detail points are calculated based on the Sobel operator. The formula is: ， ， in, is the position of the pixel value, , are the weights of the horizontal and vertical Sobel operators, , are the horizontal and vertical gradients of detail points; Based on the horizontal and vertical gradients of the acquired detail points, the gradient direction of the detail points is calculated using the formula: ， Based on this formula, the gradient direction of each detail point is calculated in radians. The gradient direction value is in the range [0, 2 ) within the range of .

4. According to the biometric access control fingerprint recognition method of claim 2, it is characterized in that: The distribution of the gradient directions of all the minutiae points in the fingerprint region based on the statistics of the gradient directions of all the minutiae points in the fingerprint region specifically includes: Based on the statistical gradient directions of all detail points in the fingerprint area, the intervals are divided and the gradient direction distribution is obtained through the direction histogram. The formula is: , in, is the gradient direction of the detail point, is an indicator function, indicating whether the gradient direction of the detail point falls within the current direction range; Based on the obtained directional histogram, normalization is performed to ensure that the sum of the directional histogram is 1. The formula is: ， Based on the normalized direction histogram, the gradient direction distribution of all detail points is obtained.

5. According to the biometric access control fingerprint recognition method of claim 2, it is characterized in that: The interpolation operation based on the gradient direction distribution to obtain the direction distribution function specifically includes: Based on the acquired direction histogram, the direction distribution is smoothed by linear interpolation. Assume that in the interval [ ) is a linear function, then the directional distribution function formula after interpolation is: , Based on the periodicity of the gradient direction of the detail points, the interpolation function is periodically extended so that exist and The values ​​at are the same.

6. According to the biometric access control fingerprint recognition method of claim 1, it is characterized in that: The adding of statistically significant descriptors to detail points specifically includes: Select the descriptor extraction area of ​​a certain detail point; Determine the four quadrants of the region based on the descriptor extraction region and the direction of the detail point; Calculate the gradient direction of all detail points in the four quadrants of the region and perform interpolation operations; The gradient distribution of the four quadrants of the region after the interpolation operation is statistically calculated to form the descriptor of the detail point, and the above steps are repeated until all detail points of the input fingerprint image are traversed.

7. The fingerprint recognition method for access control verification based on biometrics according to claim 6 is characterized in that: The method of extracting the region based on the descriptor and determining the four quadrants of the region based on the direction of the detail point specifically includes: Based on the circular extraction area, the direction of the detail point is taken as the positive direction of the horizontal axis, the detail point is taken as the origin, and then the circular area is divided into 4 quadrants according to the plane rectangular coordinate system. The rotation equation is: , Among them, [x, y] is the original coordinate, [x', y'] is the rotated coordinate, and θ is the direction of the detail point.

8. According to the biometric access control fingerprint recognition method of claim 1, it is characterized in that: The triangular vertex directions of extracting detail points specifically include: Based on the gradient direction of the acquired detail points, a triangle model is constructed, and the angles of each branch line segment are calculated to determine the main direction. The formula is: , in, are the directions of the three branches, and their average is taken to represent the main direction of the triangle vertices.

9. The fingerprint recognition method for access control verification based on biometrics technology according to claim 1 is characterized in that: The fingerprint matching based on the extracted minutiae points to obtain the fingerprint recognition result specifically includes: By calculating the distance and direction similarity between detail points, the similarity of detail points is obtained. The formula is: ， in, , For two details, is the Euclidean distance between two minutiae points, is the directional similarity of two detail points; When the similarity of the minutiae points meets a preset first threshold, adjusting the fingerprint pattern; The similarity of overlapping regions is calculated based on the normalized cross-correlation coefficient, and the formula is: ， in, is the overlap region similarity, is the overlapping area, For two fingerprint images at position The pixel value of When the similarity of the overlapping areas meets a preset second threshold, it is determined that the match is successful.

Images