A fingerprint recognition method, device and system for intelligent access control

Abstract

The present invention relates to the field of fingerprint recognition technology, and specifically to a fingerprint recognition method, device and system for intelligent access control. First, the fingerprint image to be recognized and the comparison fingerprint image entered into the database of the intelligent access control are divided into image blocks to accurately capture the local features of the fingerprint. Subsequently, the target block in the fingerprint image to be recognized and the comparison block in the comparison fingerprint image are feature matched to preliminarily determine the similar blocks. Taking into account different pressing conditions and angle deviations, the matching relationship is comprehensively evaluated by analyzing the similarity of the neighborhood blocks between the target block and the corresponding similar block, and the relative position similarity value is calculated, thereby improving the accuracy and reliability of fingerprint recognition. Finally, the relative position similarity values ​​of all target blocks and corresponding blocks are comprehensively considered to obtain a matching degree value, thereby achieving accurate fingerprint matching and user identity judgment. Even in the case of incomplete fingerprint images, the reliability of recognition can be improved.

Description

Technical Field

[0001] The present invention relates to the field of fingerprint recognition technology, and in particular to a fingerprint recognition method, device and system for intelligent access control. Background Art

[0002] With the rapid development of science and technology and people's increasing demand for security, smart access control systems, as an important means to protect public safety and personal privacy, have been widely used and deeply studied. In smart access control systems, fingerprint recognition technology has become one of the most common identity authentication methods due to its high accuracy, high reliability and ease of use.

[0003] In the fingerprint recognition process of smart access control, the existing technology usually determines whether there is a match by comparing the position of the fingerprint to be recognized with the feature points of the fingerprints already entered in the database. However, due to changes in the pressure and direction of the finger, the image of the fingerprint to be recognized in the fingerprint recognition area of ​​the smart access control will be incomplete. Therefore, based on the incomplete image, only the position of the feature points is compared, which will reduce the recognition reliability, resulting in low fingerprint matching accuracy and fingerprint recognition failure. Summary of the invention

[0004] In order to solve the technical problem that only position comparison of feature points on the basis of incomplete images will lead to reduced recognition reliability, low fingerprint matching accuracy, and fingerprint recognition failure, the purpose of the present invention is to provide a fingerprint recognition method, device and system for intelligent access control, and the technical solutions adopted are as follows:

[0005] Obtain the fingerprint image to be identified in the smart access control fingerprint identification area; obtain the comparison fingerprint image pre-entered in the smart access control database;

[0006] Dividing the fingerprint image to be identified to obtain target blocks, dividing the comparison fingerprint image to obtain comparison blocks; performing feature matching on the target blocks and the comparison blocks to obtain similar blocks corresponding to the target blocks;

[0007] Determine the neighborhood blocks of each target block in the fingerprint image to be identified, and determine the neighborhood blocks of each comparison block in the comparison fingerprint image; analyze the similarity of the neighborhood blocks between each target block and the corresponding similar block, and determine the relative position similarity value between the target block and the corresponding similar block;

[0008] Based on the relative position similarity values ​​between all target blocks and corresponding similar blocks, the matching degree value between the fingerprint image to be identified and the comparison fingerprint image is determined; and fingerprint identification is performed based on the matching degree value.

[0009] Furthermore, the method for obtaining similar blocks includes:

[0010] Based on the scale-invariant feature transformation algorithm, feature points in the target block and the comparison block and feature values ​​corresponding to the feature points are respectively obtained; the difference between the feature values ​​of the feature points between the target block and the comparison block is compared with a preset difference threshold to determine the matching feature point pair;

[0011] Determine the similarity value between the target block and the comparison block based on the positional relationship between the feature point in the matching feature point pair and other feature points, and the number of matching feature point pairs;

[0012] For any target block, among the similarity values ​​greater than a preset similarity threshold, the comparison block corresponding to the maximum similarity value is taken as the similar block corresponding to the target block.

[0013] Furthermore, the method for obtaining the similarity value includes:

[0014] In each set of matching feature point pairs corresponding to any target block and any comparison block, the feature point in the target block is taken as the first starting point, the feature point in the fingerprint image to be identified that is closest to the first starting point is taken as the first end point, a first direction vector is obtained, the angle value between the first direction vector and the preset direction is taken as the first angle value, the feature point in the comparison block is taken as the second starting point, the feature point in the comparison fingerprint image that is closest to the second starting point is taken as the second end point, a second direction vector is obtained, and the angle value between the second direction vector and the preset direction is taken as the second angle value; based on the difference between the first angle value and the second angle value corresponding to all matching feature point pairs in the target block and the comparison block, a first difference factor between the target block and the comparison block is determined;

[0015] taking the difference between the number of feature points in the target block and the number of matching feature point pairs between the target block and the comparison block as a second difference factor;

[0016] Based on the first difference factor and the second difference factor, a similarity value between the target block and the comparison block is determined, and both the first difference factor and the second difference factor are negatively correlated with the similarity value.

[0017] Furthermore, the method for obtaining the relative position similarity value includes:

[0018] Marking the consistency between the neighborhood blocks of the target block and the neighborhood blocks of the similar blocks corresponding to the target block to obtain a marking sequence corresponding to the target block, and determining the position similarity factor between each target block and the corresponding similar block based on the numerical features in the marking sequence;

[0019] Based on the corresponding value of each neighboring block of the target block in the tag sequence, the position similarity factor between the target block and the similar block is adjusted to obtain the influence factor of each neighboring block of the target block on the target block;

[0020] The relative position similarity value between the target block and the corresponding similar block is obtained by combining the position similarity factor between the target block and the corresponding similar block and the influence factor of all the neighboring blocks of all the target blocks on the target block, and the position similarity factor and the influence factor are both positively correlated with the relative position similarity value.

[0021] Furthermore, the method for obtaining the location similarity factor includes:

[0022] In the fingerprint image to be identified, the neighborhood block of the target block is used as the reference block; in the fingerprint image to be compared, the neighborhood block of the target block corresponding to the similar block is used as the analysis block;

[0023] In the analysis block, if there is no analysis block similar to the reference block, all the reference blocks corresponding to the target block are marked as 0, and the marking sequence corresponding to the target block is obtained;

[0024] If there is an analysis block similar to the reference block, any group of similar analysis blocks and reference blocks are used as the starting block. Starting from the starting block, the similar block of the reference block in the comparison image is judged in a clockwise direction to see whether it is consistent with the analysis block at the corresponding position. If they are consistent, the reference block is marked as 1, otherwise it is marked as 0. The marking sequence corresponding to the target block is obtained.

[0025] In the tag sequence corresponding to the target block, the position similarity factor between the target block and the corresponding similar block is determined based on the difference between the number of values ​​1 and the total number of values ​​in the tag sequence.

[0026] Furthermore, the method for obtaining the impact factor includes:

[0027] The product of the numerical value corresponding to each neighboring block of the target block in the tag sequence and the position similarity factor of the target block is used as the influence factor of the neighboring block of the target block on the target block.

[0028] Furthermore, the method for obtaining the matching degree value includes:

[0029] In the fingerprint image to be identified, the ratio of the number of target blocks whose relative position similarity values ​​are greater than a preset position similarity threshold to the total number of all target blocks is used as a matching degree value.

[0030] Further, the fingerprint recognition based on the matching degree value includes:

[0031] The matching degree values ​​of the fingerprint image to be identified and all the comparison fingerprint images are compared with the preset matching threshold. When there is a matching degree value greater than the preset matching threshold, the fingerprint recognition is judged to be successful, otherwise the fingerprint recognition is judged to be failed.

[0032] The present invention also proposes a fingerprint recognition system for intelligent access control, the system comprising:

[0033] The image acquisition module is used to acquire the fingerprint image to be identified in the intelligent access control fingerprint recognition area; and acquire the comparison fingerprint image pre-entered in the intelligent access control database;

[0034] A feature matching module is used to divide the fingerprint image to be identified to obtain a target block, divide the comparison fingerprint image to obtain a comparison block; perform feature matching on the target block and the comparison block to obtain a similar block corresponding to the target block;

[0035] A relative position analysis module is used to determine the neighborhood blocks of each target block in the fingerprint image to be identified, and to determine the neighborhood blocks of each comparison block in the comparison fingerprint image; analyze the similarity of the neighborhood blocks between each target block and the corresponding similar block, and determine the relative position similarity value between the target block and the corresponding similar block;

[0036] The fingerprint matching module is used to determine the matching degree value between the fingerprint image to be identified and the comparison fingerprint image based on the relative position similarity value between all target blocks and corresponding similar blocks; and perform fingerprint identification based on the matching degree value.

[0037] The present invention also proposes a fingerprint recognition device for smart access control, the device comprising a processor and a memory, the memory storing at least one instruction, at least one program, code set or instruction set, and at least one instruction, at least one program, code set or instruction set, when loaded and executed by the processor, implements the steps of any one of the methods.

[0038] The present invention has the following beneficial effects:

[0039] First, the fingerprint image to be identified and the comparison fingerprint image entered in the database are obtained, and the fingerprint image to be identified and the comparison fingerprint image are divided into image blocks respectively to obtain the target block and the comparison block, which can capture the local features of the fingerprint more accurately. Then, the target block and the comparison block are feature matched, and the similarity between the target block and the comparison block can be analyzed, so as to preliminarily determine the similar block of the target block. Due to different finger pressing strengths and pressing directions, the fingerprint image to be identified will be incomplete and have angle deviations compared with the comparison fingerprint image, but the relative position relationship between the various features of the fingerprint will not change due to the occurrence of the above situation. Therefore, in the present invention, by analyzing the similarity between the neighborhood block of the target block and the neighborhood block of the corresponding similar block of the target block, the matching relationship between the target block and the corresponding similar block can be more comprehensively evaluated, and then the relative position similarity value between the target block and the corresponding similar block is determined, thereby improving the accuracy and reliability of fingerprint recognition. Finally, by comprehensively considering the relative position similarity values ​​between all target blocks and corresponding blocks in the image, the matching degree value between the fingerprint image to be identified and the comparison fingerprint image is obtained, and fingerprint recognition is performed based on the matching degree value. Therefore, even if the fingerprint image is incomplete, fingerprint matching can be performed accurately, which improves the reliability of recognition and more accurately determines the user identity. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. 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 paying creative work.

[0041] Figure 1 A method flow chart of a fingerprint recognition method for intelligent access control provided by one embodiment of the present invention;

[0042] Figure 2 A schematic diagram of comparing fingerprint images provided by an embodiment of the present invention;

[0043] Figure 3 A schematic diagram of a fingerprint pattern to be identified provided by an embodiment of the present invention;

[0044] Figure 4 A method flow chart of a method for obtaining similar blocks provided by one embodiment of the present invention;

[0045] Figure 5 A method flow chart of a method for obtaining a relative position similarity value provided by one embodiment of the present invention;

[0046] Figure 6 A system block diagram of a fingerprint recognition system for intelligent access control provided by an embodiment of the present invention;

[0047] Figure 7 A schematic diagram of a fingerprint recognition device for smart access control provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0048] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of the fingerprint recognition method, device and system for intelligent access control proposed by the present invention, its specific implementation method, structure, features and effects, in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0049] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0050] The following is a detailed description of a fingerprint recognition method for smart access control provided by the present invention in conjunction with the accompanying drawings.

[0051] See also Figure 1 , which shows a method flow chart of a fingerprint recognition method for smart access control provided by an embodiment of the present invention, the method comprising the following steps:

[0052] Step S1: Obtain a fingerprint image to be identified in the smart access control fingerprint identification area; obtain a comparison fingerprint image pre-entered in the smart access control database.

[0053] Fingerprint recognition in smart access control systems is a common and effective authentication technology. By reading and analyzing the biometric features of individual fingerprints and comparing them with the biometric information in the database, the identity is confirmed and access to specific areas or equipment is controlled.

[0054] In this process, when the user places his finger on the smart access control fingerprint recognition area, the access control system can start the image acquisition device, such as a high-resolution camera or fingerprint scanner, to quickly capture the user's fingerprint image; at the same time, it is also necessary to obtain the fingerprint image pre-entered in the smart access control database. After obtaining the user's fingerprint image and the fingerprint image entered in the database, they are grayed and binarized to facilitate the subsequent extraction of fingerprint image features, and then the Canny algorithm is used to perform edge detection to obtain the texture image corresponding to the fingerprint image. Finally, the texture image of the fingerprint on the access control fingerprint recognition area is used as the fingerprint image to be identified, and the texture image corresponding to the fingerprint in the database is used as the comparison fingerprint image.

[0055] It should be noted that the fingerprint image acquisition device can be adjusted according to the implementation scenario and is not limited here; the operations of graying and binarizing the image and the Canny algorithm are technical means well known to those skilled in the art and will not be described in detail here.

[0056] Step S2: Divide the fingerprint image to be identified to obtain target blocks, divide the comparison fingerprint image to obtain comparison blocks; perform feature matching on the target blocks and the comparison blocks to obtain similar blocks corresponding to the target blocks.

[0057] Since the fingerprints entered into the smart access control database are usually relatively complete, please refer to Figure 2 , which shows a schematic diagram of comparing fingerprint images in one embodiment of the present invention; however, in actual use by users, due to changes in the pressing force and pressing direction of the finger, the fingerprint image obtained in the smart access control fingerprint recognition area may be incomplete, and may only contain part of the fingerprint. Please participate Figure 3 , which shows a schematic diagram of a fingerprint image to be identified in an embodiment of the present invention.

[0058] Therefore, in order to more accurately match images, a strategy from local to global can be adopted. The fingerprint image to be identified is divided to obtain multiple target blocks. At the same time, the comparison fingerprint image is divided in the same way to obtain multiple comparison blocks. This division can ensure that the local details in the fingerprint image can be focused on, thereby capturing the local features of the fingerprint more comprehensively.

[0059] It should be noted that the image division method may specifically be to evenly divide the image into a preset number of image blocks, where the preset number is set to 32. The specific value may be adjusted according to the implementation scenario and is not limited here.

[0060] Then, feature matching is performed on the target block and the comparison block one by one to determine similar image blocks.

[0061] Preferably, in one embodiment of the present invention, performing feature matching on the target block and the comparison block to obtain a similar block corresponding to the target block includes:

[0062] See also Figure 4 , which shows a method flow chart of a method for obtaining similar blocks in one embodiment of the present invention, the method comprising the following steps:

[0063] Step S201: Acquire feature points in the target block and the comparison block and match the feature points to determine matching feature point pairs.

[0064] The scale-invariant feature transform (SIFT) algorithm can extract feature points with scale invariance and rotation invariance, so that even if the fingerprint image to be identified is slightly deformed or rotated, the corresponding feature points can be accurately matched; at the same time, for each feature point, the local gradient of the image is measured in its surrounding neighborhood, and the gradient information is converted into a feature vector, that is, the eigenvalue (descriptor) corresponding to each feature point is obtained; therefore, the feature points in each target block and each comparison block and the eigenvalues ​​corresponding to the feature points are obtained based on the SIFT algorithm.

[0065] Then, for any target block and any comparison block, the characteristic values ​​between the feature points in the target block and the feature points in the comparison block can be calculated for the Euclidean distance. At this time, each feature point in the target block has a difference index with all feature points in the comparison block, and the smaller the difference index, the closer the two feature points are in the feature space, which means that the two feature points may be the representation of the same feature point in different images. Therefore, the minimum value of all difference indexes of each feature point in the target block is compared with the preset difference threshold. When the minimum difference index is less than the preset difference threshold, the feature point corresponding to the minimum difference index in the comparison block is used as the matching point of the feature point in the target block. According to the above method, the matching feature point pair between the target block and the comparison block can be obtained.

[0066] It should be noted that the preset difference threshold is set to 8, and the specific value can be adjusted according to the implementation scenario and is not limited here.

[0067] The fork in the fingerprint texture is an important feature, which provides very valuable information for fingerprint recognition technology. In the fingerprint image, when the direction of the fingerprint texture suddenly changes from one direction to another, a fork is formed. These forks usually appear as corner points. Therefore, in other embodiments of the present invention, the Harris corner point detection technology can be used to obtain the corner points in the target block and the comparison block, and the corner points can be used as feature points. Then, based on the SIFT algorithm, the feature value of the feature point is generated by comparing the gradient information around the feature point; thereby continuing to judge the matching feature point pairs.

[0068] It should be noted that the Harris corner detection technology is a technical means well known to those skilled in the art and will not be described in detail here.

[0069] Step S202: Determine the similarity value between the target block and the comparison block based on the positional relationship between the feature point in the matching feature point pair and other feature points, and the number of matching feature point pairs.

[0070] When two image blocks are highly similar, the spatial distribution relationship of the matching feature points in the image blocks should be relatively consistent. Therefore, in each set of matching feature point pairs, the position distribution characteristics of each feature point in the image block to which it belongs are analyzed, which can be specifically characterized by the position distance between it and other feature points.

[0071] In each set of matching feature point pairs of any target block and any comparison block, the feature point in the target block is taken as the first starting point, and the feature point in the fingerprint image to be identified that is closest to the first starting point is taken as the first end point to obtain a first direction vector, and the angle value between the first direction vector and the preset direction is taken as the first angle value; the feature point in the comparison block is taken as the second starting point, and the feature point in the comparison fingerprint image that is closest to the second starting point is taken as the second end point to obtain a second direction vector, and the angle value between the second direction vector and the preset direction is taken as the second angle value.

[0072] The first angle value can reflect the position distribution of the feature points in the target block, and the second angle value can reflect the position distribution of the feature points in the comparison block. Therefore, by analyzing the difference between the first angle value and the second angle value corresponding to all matching feature point pairs in the target block and the comparison block, the first difference factor between the target block and the comparison block can be determined.

[0073] Then, the difference between the number of feature points in the target block and the number of matching feature point pairs between the target block and the comparison block is used as the second difference factor.

[0074] Finally, based on the first difference factor and the second difference factor, the similarity value between the target block and the comparison block is determined, and the first difference factor and the second difference factor are both negatively correlated with the similarity value. The formula model of the similarity value can be specifically, for example:

[0075]

[0076] Among them, XC i,j Represents the similarity value between the i-th target block and the j-th comparison block; M i Represents the number of feature points in the i-th target block; N i,j represents the number of matching feature point pairs between the i-th target block and the j-th comparison block; θ i,k represents the first angle value corresponding to the feature point in the i-th target block in the k-th matching feature point pair; θ j,k It represents the second angle value corresponding to the feature point in the jth comparison block in the kth matching feature point pair; exp[ ] represents an exponential function with the natural constant e as the base.

[0077] In the formula model of the similarity value, the difference between the first angle value and the second angle value of each matching feature point pair in the target block and the comparison block is calculated. i,k -θ j,k |, when the difference value is larger, it means that in the matching feature point pair, the similarity between the spatial distribution features of the two feature points is lower. On the contrary, when the difference value is smaller, it means that in the matching feature point pair, the similarity between the spatial distribution features of the two feature points is higher. The difference values ​​of all matching feature point pairs in the target block and the comparison block are averaged to obtain the first difference factor between the target block and the comparison block. It can represent the average level of the difference. The larger the value of the first difference factor is, the lower the similarity between the target block and the comparison block is. Since the difference threshold is set to limit the matching degree between the feature points when obtaining the matching feature point pairs, the more the number of successfully matched feature points is, the higher the similarity between the target block and the comparison block is. Therefore, the difference between the number of feature points in the target block and the number of matching feature point pairs between the target block and the comparison block is calculated as the second difference factor (M i -N i,j ), when the second difference factor value is smaller, it means that the similarity between the target block and the comparison block is higher, otherwise, it is considered that the similarity between the target block and the comparison block is lower. Based on the above analysis, the product of the first difference factor and the second difference factor is negatively correlated to achieve logical relationship correction, thereby obtaining the similarity value between the target block and the comparison block, and the larger the similarity value, the more similar the features between the target block and the comparison block.

[0078] It should be noted that the preset direction is set to the horizontal right direction, and the setting of the specific direction can be adjusted according to the implementation scenario and is not limited here.

[0079] Step S203: Based on the similarity value between the target block and the comparison block, determine the similar block corresponding to each target block.

[0080] Based on the above steps, for any target block in the fingerprint image to be identified, there will be a similarity value between it and each comparison block in the comparison fingerprint image. Therefore, for any target block, among the similarity values ​​greater than the preset similarity threshold, the comparison block corresponding to the largest similarity value is used as the similar block corresponding to the target block.

[0081] It should be noted that the similarity threshold is set to 0.6, and the specific value can be adjusted according to the implementation scenario and is not limited here.

[0082] So far, through the above process, the similar block corresponding to each target block in the fingerprint image to be identified can be obtained.

[0083] Step S3: Determine the neighborhood blocks of each target block in the fingerprint image to be identified, and determine the neighborhood blocks of each comparison block in the comparison fingerprint image; analyze the similarity of the neighborhood blocks between each target block and the corresponding similar block, and determine the relative position similarity value between the target block and the corresponding similar block.

[0084] Due to different finger pressure strength and pressing direction, the fingerprint image to be identified may be incomplete and have angle deviations compared with the comparison fingerprint image. This may cause the position and direction of the same texture in the fingerprint image to be identified and the comparison fingerprint image to change. However, the relative position relationship between the various features of the fingerprint will not change due to the occurrence of the above situation. Therefore, in an embodiment of the present invention, by analyzing the similarity between the neighborhood blocks of the target block and the neighborhood blocks of the similar blocks corresponding to the target block, the relative position similarity value between the target block and the corresponding similar blocks is determined, thereby providing data support for the subsequent analysis of the degree of matching between the fingerprint image to be identified and the comparison fingerprint image.

[0085] First, it is necessary to determine the neighborhood blocks of each target block in the fingerprint image to be identified and the neighborhood blocks of each comparison block in the comparison fingerprint image.

[0086] Then, the similarity between the neighborhood blocks of the target block and the neighborhood blocks of the similar blocks corresponding to the target block can be analyzed, so as to determine the relative position similarity value between the target block and the corresponding similar blocks.

[0087] Preferably, in one embodiment of the present invention, the method for obtaining the relative position similarity value includes:

[0088] See also Figure 5, which shows a method flow chart of a method for obtaining a relative position similarity value in one embodiment of the present invention, the method comprising the following steps:

[0089] Step S301: Mark the consistency between the neighborhood blocks of the target block and the neighborhood blocks of the similar blocks to obtain a marking sequence corresponding to the target block, and determine the position similarity factor between the target block and the similar blocks based on the numerical features in the marking sequence.

[0090] In the fingerprint image to be identified, the neighborhood block of the target block is used as the reference block, and in the fingerprint image to be compared, the neighborhood block of the target block corresponding to the similar block is used as the analysis block.

[0091] Since it is impossible to determine the angle relationship between the fingerprint image to be identified and the comparison fingerprint image at this time, it is possible to determine whether there is a similar block to the reference block in the analysis block. If there is no analysis block similar to the reference block, all reference blocks corresponding to the target block are marked as 0 to obtain the marking sequence corresponding to the target block.

[0092] If there is an analysis block similar to the reference block, any group of similar analysis blocks and reference blocks are used as the starting block. Then, starting from the starting block, in a clockwise direction, determine whether the similar block of the reference block in the comparison image is consistent with the analysis block at the corresponding position. If they are consistent, the reference block is marked as 1, otherwise it is marked as 0; thus, the label sequence corresponding to the target block is obtained.

[0093] It should be noted that in this embodiment of the present invention, the neighborhood blocks are set as four neighborhoods, that is, the four adjacent image blocks above, below, left and right of each target block are used as its neighborhood blocks, and similarly, the four adjacent image blocks above, below, left and right of each comparison block are used as its neighborhood blocks. Here, an example is given to illustrate the process of obtaining the marking sequence: for any target block in the comparison image, it is recorded as target block 1, and the similar block corresponding to the target block 1 in the comparison fingerprint image is recorded as similar block a. The neighborhood block above the target block 1 is recorded as reference block 2, and in a clockwise direction, it is recorded as reference block 3, reference block 4, and reference block 5; similarly, the neighborhood block above the similar block a is recorded as analysis block b, and in a clockwise direction, it is recorded as analysis block c, analysis block d, and analysis block e. First, determine whether the reference block 2, reference block 3, reference block 4, and reference block 5 are similar to the analysis block b, analysis block c, analysis block d, and analysis block e. If not, then all four reference blocks corresponding to the target block 1 are marked as 0, and the marking sequence corresponding to the target block 1 is [0, 0, 0, 0]. If there is a reference block 2 that is similar to the analysis block c, and a reference block 4 that is similar to the analysis block e, then any one of these groups is selected as the starting block. Here, select reference block 2 and analysis block c, and then compare whether the similar block of reference block 2 in the comparison image is consistent with the analysis block c, that is, analyze whether they are the same block. If they are consistent, then mark reference block 2 as 1, and continue to compare whether the similar block of reference block 3 is consistent with the analysis block d. If they are not consistent, then mark reference block 3 as 0, and continue to compare whether the similar block of reference block 4 is consistent with the analysis block e. If they are consistent, mark reference block 4 as 1, and then whether the similar block of reference block 5 is consistent with the analysis block b. If they are not consistent, then mark reference block 5 as 0. At this time, the comparison sequence corresponding to the target block 1 is [1, 0, 1, 0]. In other embodiments of the present invention, the number of neighborhood blocks is not limited and may also be set to 8 neighborhoods. The specific number may be adjusted according to the implementation scenario and is not limited here.

[0094] Finally, in the tag sequence corresponding to the target block, the proportion of the number of values ​​1 is used as the position similarity factor between the target block and the corresponding similar block. The formula model of the position similarity factor includes:

[0095]

[0096] Wherein, WX represents the position similarity factor between each target block and the corresponding similar block; n represents the number of values ​​1 in the tag sequence corresponding to each target block; and w represents the total number of values ​​in the tag sequence corresponding to each target block.

[0097] In the formula model of the position similarity factor, if the number of values ​​​​of 1 in the label sequence is greater, the ratio The larger it is, the higher the consistency between the similar blocks of the neighborhood blocks of the target block and the neighborhood blocks of the corresponding similar blocks of the target block is, that is, based on the analysis of the similarity between the neighborhood blocks, the higher the similarity between the target block and the corresponding similar blocks is.

[0098] In other embodiments of the present invention, the value after negative correlation mapping between the number of values ​​1 and the total number of values ​​in the tag sequence corresponding to the target block can be used as the position similarity factor between the target block and the corresponding similar block. The formula model of the position similarity factor includes:

[0099] WX=exp[-|nw|]

[0100] Where WX represents the position similarity factor between each target block and the corresponding similar block; n represents the number of values ​​1 in the tag sequence corresponding to each target block; w represents the total number of values ​​in the tag sequence corresponding to each target block; exp[ ] represents an exponential function with the natural constant e as the base.

[0101] In the formula model of the position similarity factor, if the number of values ​​1 in the mark sequence is more, the difference value |nw| is smaller, which means that the similarity block of the neighborhood block of the target block is more consistent with the neighborhood block of the corresponding similar block of the target block, that is, on the basis of analyzing the similarity between the neighborhood blocks, the similarity between the target block and the corresponding similar block is higher, so the difference value is negatively correlated and normalized to achieve logical relationship correction, thereby obtaining the position similarity factor between the target block and the corresponding similar block.

[0102] Step S302: Based on the corresponding value of each neighboring block of the target block in the mark sequence, the position similarity factor between the target block and the similar block is adjusted to obtain the influence factor of each neighboring block of the target block on the target block.

[0103] In the tag sequence corresponding to the target block, different values ​​can reflect the matching degree between the target block and the corresponding similar block by the neighborhood blocks of the target block, so here, the influence of each neighborhood block of the target block on the position similarity factor of the target block is analyzed.

[0104] The product of the value corresponding to each neighboring block of the target block in the mark sequence and the position similarity factor of the target block is used as the influence factor of the neighboring block of the target block on the target block. The formula model of the influence factor includes:

[0105] YX p =WX×B p

[0106] Among them, YX prepresents the influence factor of the pth neighboring block of each target block on the target block; WX represents the position similarity factor between each target block and the corresponding similar block; B p Represents the value corresponding to the pth neighboring block of each target block in the label sequence.

[0107] In the formula model of the influence factor, when the corresponding value of the neighborhood block of the target block in the mark sequence is 0, it is considered that the neighborhood block has no reference value for the matching between the target block and the corresponding similar block, so the value of the influence factor is 0; conversely, if the corresponding value of the neighborhood block of the target block in the mark sequence is 1, it is considered that the neighborhood block has a reference value for the matching between the target block and the corresponding similar block, so the value of the influence factor is 1.

[0108] Based on this step, the influence factors of all neighboring blocks corresponding to each target block can be obtained.

[0109] Step S303: The position similarity factor between the target block and the corresponding similar block and the influence factor of all neighboring blocks on the target block are integrated to obtain the relative position similarity value between the target block and the similar block.

[0110] Based on the analysis in step S301 and step S302, it can be known that the higher the similarity between the target block and the corresponding similar block, the larger the quantized position similarity factor will be; the influence factor corresponding to the neighborhood block of the target block is 1, indicating that the neighborhood block has a reference value for the matching between the target block and the corresponding similar block. Therefore, by combining the position similarity factor and the influence factor, the relative position similarity value between the target block and the similar block can be obtained, and the position similarity factor and the influence factor are both positively correlated with the relative position similarity value. The formula model of the relative position similarity value can be, for example, as follows:

[0111]

[0112] Among them, XZ represents the relative position similarity value between each target block and the corresponding similar block; WX represents the position similarity factor between each target block and the corresponding similar block; YX p It represents the influence factor of the pth neighboring block of each target block on the target block; norm[] represents the normalization function.

[0113] In the formula model of relative position similarity value, the influence factors corresponding to all neighboring blocks of the target block are integrated to obtain a comprehensive influence value The larger the value is, the higher the reference value of the neighborhood blocks around the target block for judging the position similarity between the target block and its corresponding similar block, and the more reliable the calculated result is. Therefore, the comprehensive influence value and the position similarity factor between the target block and the corresponding similar block are combined, and the sum of the two is normalized to obtain the relative position similarity value between the target block and the corresponding similar block. The larger the value is, the more consistent the position of the target block in the fingerprint image to be identified is with the position of the corresponding similar block in the comparison fingerprint image, and it is considered that the consistency between the target block and the corresponding similar block is higher.

[0114] So far, the relative position similarity values ​​between all target blocks and corresponding similar blocks can be obtained through the above process.

[0115] It should be noted that if a target block does not have a corresponding similar block, the relative position similarity value corresponding to the target block is set to a preset value, and since the relative position similarity value ranges from 0 to 1, the preset value is set to 0 here.

[0116] Step S4: Based on the relative position similarity values ​​between all target blocks and corresponding similar blocks, determine the matching degree value between the fingerprint image to be identified and the comparison fingerprint image; perform fingerprint identification based on the matching degree value.

[0117] In the process of fingerprint recognition, the matching depends not only on the similarity between the image blocks, but also on the relative position relationship of these image blocks in the fingerprint image. This relative position relationship is crucial to ensure the accuracy of fingerprint recognition, because the relative position relationship between the features of the fingerprint usually does not change. Therefore, after accurately measuring the similarity of the image blocks in the two fingerprint images in terms of relative position relationship, a matching degree value can be determined based on the relative position similarity value between all target blocks and the corresponding similar blocks. This matching degree value is a quantitative indicator used to indicate the similarity between the fingerprint image to be identified and the comparison fingerprint image. Finally, fingerprint recognition is performed based on this matching degree value.

[0118] Preferably, in one embodiment of the present invention, the method for obtaining the matching degree value includes:

[0119] For any comparison fingerprint image, each target block in the fingerprint image to be identified corresponds to a relative position similarity value, and the larger the relative position similarity value, the more similar blocks there are in the comparison fingerprint image, which are in the same relative position as the target blocks in the fingerprint image to be identified. Therefore, the ratio of the number of target blocks whose relative position similarity values ​​are greater than the preset position similarity threshold to the total number of all target blocks is used as the matching degree value. At this time, the larger the matching degree value, the higher the similarity between the fingerprint image to be identified and the comparison fingerprint image.

[0120] It should be noted that the location similarity threshold is set to 0.5, and the specific value can be adjusted according to the implementation scenario and is not limited here.

[0121] Based on the above process, the matching degree value between the fingerprint image to be identified and all the comparison fingerprint images pre-entered in the database of the intelligent access control system can be obtained, and then fingerprint identification can be performed based on the matching degree value.

[0122] Preferably, in one embodiment of the present invention, fingerprint recognition is performed based on the matching degree value, including:

[0123] Since the larger the matching degree value, the higher the similarity between the fingerprint image to be identified and the comparison fingerprint image, the matching degree value between the fingerprint image to be identified and all the comparison fingerprint images is compared with the preset matching threshold. When there is a matching degree value greater than the preset matching threshold, it is considered that there is a comparison fingerprint image belonging to the same fingerprint as the fingerprint image to be identified in the database of the smart access control system, and the fingerprint recognition is judged to be successful; otherwise, the fingerprint recognition is judged to have failed.

[0124] It should be noted that the matching threshold is set to 0.8, and the specific value can be adjusted according to the implementation scenario and is not limited here; in order to facilitate the calculation, all indicator data involved in the calculation in the embodiment of the present invention are pre-processed to eliminate the dimension effect. The specific means of removing the dimension effect are technical means well known to those skilled in the art and are not limited here.

[0125] In summary, firstly, the fingerprint image to be identified and the comparative fingerprint image entered into the database are obtained, and the fingerprint image to be identified and the comparative fingerprint image are divided into image blocks respectively to obtain the target block and the comparative block, so as to more accurately capture the local features of the fingerprint. Then, the target block and the comparative block are feature matched, and the similarity between the target block and the comparative block can be analyzed, so as to preliminarily determine the similar block of the target block. Due to different finger pressing strength and pressing direction, the fingerprint image to be identified will be incomplete and have angle deviations compared with the comparative fingerprint image, but the relative position relationship between the various features of the fingerprint will not change due to the occurrence of the aforementioned situation. Therefore, in the embodiment of the present invention, by analyzing the similarity between the neighborhood block of the target block and the neighborhood block of the corresponding similar block of the target block, the matching relationship between the target block and the corresponding similar block can be more comprehensively evaluated, and then the relative position similarity value between the target block and the corresponding similar block is determined, thereby improving the accuracy and reliability of fingerprint recognition. Finally, by comprehensively considering the relative position similarity values ​​between all target blocks and corresponding blocks in the image, the matching degree value between the fingerprint image to be identified and the comparison fingerprint image is obtained, and fingerprint recognition is performed based on the matching degree value. Therefore, even if the fingerprint image is incomplete, fingerprint matching can be performed accurately, which improves the reliability of recognition and more accurately determines the user identity.

[0126] The present invention also proposes a fingerprint recognition system for intelligent access control, please refer to Figure 6 , the system includes:

[0127] The image acquisition module 601 is used to acquire the fingerprint image to be identified in the smart access control fingerprint recognition area; and acquire the comparison fingerprint image pre-entered in the smart access control database;

[0128] The feature matching module 602 is used to divide the fingerprint image to be identified to obtain target blocks, divide the comparison fingerprint image to obtain comparison blocks; perform feature matching on the target blocks and the comparison blocks to obtain similar blocks corresponding to the target blocks;

[0129] The relative position analysis module 603 is used to determine the neighborhood blocks of each target block in the fingerprint image to be identified, and determine the neighborhood blocks of each comparison block in the comparison fingerprint image; analyze the similarity of the neighborhood blocks between each target block and the corresponding similar block, and determine the relative position similarity value between the target block and the corresponding similar block;

[0130] The fingerprint matching module 604 is used to determine the matching degree value between the fingerprint image to be identified and the comparison fingerprint image based on the relative position similarity values ​​between all target blocks and corresponding similar blocks; and perform fingerprint identification based on the matching degree value.

[0131] It should be noted that the system provided in the above embodiment is only illustrated by the division of the above functional modules. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the computer device is divided into different functional modules to complete all or part of the functions described above. In addition, the fingerprint recognition system for smart access control provided in the above embodiment and the fingerprint recognition method embodiment for smart access control belong to the same concept. The specific implementation process is detailed in the method embodiment and will not be repeated here.

[0132] The present invention also provides a fingerprint recognition device for intelligent access control. Figure 7 , the equipment structure includes:

[0133] Processor 700, memory 701, bus 702 and communication interface 703, wherein the processor 700, communication interface 703 and memory 701 are connected via bus 702; wherein the memory 701 may include a high-speed random access memory, the bus 702 may be an ISA bus, a PCI bus or an EISA bus, etc., and the processor 700 may be an integrated circuit chip with signal processing capabilities.

[0134] 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 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.

[0135] 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.

Claims

1. A fingerprint recognition method for intelligent access control, characterized in that: The method comprises: Obtain the fingerprint image to be identified in the smart access control fingerprint identification area; obtain the comparison fingerprint image pre-entered in the smart access control database; Dividing the fingerprint image to be identified to obtain target blocks, dividing the comparison fingerprint image to obtain comparison blocks; performing feature matching on the target blocks and the comparison blocks to obtain similar blocks corresponding to the target blocks; Determine the neighborhood blocks of each target block in the fingerprint image to be identified, and determine the neighborhood blocks of each comparison block in the comparison fingerprint image; analyze the similarity of the neighborhood blocks between each target block and the corresponding similar block, and determine the relative position similarity value between the target block and the corresponding similar block; Based on the relative position similarity values ​​between all target blocks and corresponding similar blocks, determining the matching degree value between the fingerprint image to be identified and the comparison fingerprint image; and performing fingerprint identification based on the matching degree value; The method for obtaining the relative position similarity value comprises: Marking the consistency between the neighborhood blocks of the target block and the neighborhood blocks of the similar blocks corresponding to the target block to obtain a marking sequence corresponding to the target block, and determining the position similarity factor between each target block and the corresponding similar block based on the numerical features in the marking sequence; Based on the corresponding value of each neighboring block of the target block in the tag sequence, the position similarity factor between the target block and the similar block is adjusted to obtain the influence factor of each neighboring block of the target block on the target block; The relative position similarity value between the target block and the corresponding similar block is obtained by combining the position similarity factor between the target block and the corresponding similar block and the influence factor of all the neighboring blocks of all the target blocks on the target block, and the position similarity factor and the influence factor are both positively correlated with the relative position similarity value; The method for obtaining the position similarity factor includes: In the fingerprint image to be identified, the neighborhood block of the target block is used as the reference block; in the fingerprint image to be compared, the neighborhood block of the target block corresponding to the similar block is used as the analysis block; In the analysis block, if there is no analysis block similar to the reference block, all the reference blocks corresponding to the target block are marked as 0, and the marking sequence corresponding to the target block is obtained; If there is an analysis block similar to the reference block, any group of similar analysis blocks and reference blocks are used as the starting block. Starting from the starting block, the similar block of the reference block in the comparison image is judged in a clockwise direction to see whether it is consistent with the analysis block at the corresponding position. If they are consistent, the reference block is marked as 1, otherwise it is marked as 0. The marking sequence corresponding to the target block is obtained. In the tag sequence corresponding to the target block, based on the difference between the number of values ​​1 and the total number of values ​​in the tag sequence, determine a position similarity factor between the target block and the corresponding similar block; The method for obtaining the impact factor includes: The product of the numerical value corresponding to each neighboring block of the target block in the tag sequence and the position similarity factor of the target block is used as the influence factor of the neighboring block of the target block on the target block.

2. According to claim 1, a fingerprint recognition method for intelligent access control is characterized in that: The method for obtaining the similar blocks includes: Based on the scale-invariant feature transformation algorithm, feature points in the target block and the comparison block and feature values ​​corresponding to the feature points are respectively obtained; the difference between the feature values ​​of the feature points between the target block and the comparison block is compared with a preset difference threshold to determine the matching feature point pair; Determine the similarity value between the target block and the comparison block based on the positional relationship between the feature point in the matching feature point pair and other feature points, and the number of matching feature point pairs; For any target block, among the similarity values ​​greater than a preset similarity threshold, the comparison block corresponding to the maximum similarity value is taken as the similar block corresponding to the target block.

3. A fingerprint recognition method for intelligent access control according to claim 2, characterized in that: The method for obtaining the similarity value includes: In each set of matching feature point pairs corresponding to any target block and any comparison block, the feature point in the target block is taken as the first starting point, the feature point in the fingerprint image to be identified that is closest to the first starting point is taken as the first end point, a first direction vector is obtained, the angle value between the first direction vector and the preset direction is taken as the first angle value, the feature point in the comparison block is taken as the second starting point, the feature point in the comparison fingerprint image that is closest to the second starting point is taken as the second end point, a second direction vector is obtained, and the angle value between the second direction vector and the preset direction is taken as the second angle value; based on the difference between the first angle value and the second angle value corresponding to all matching feature point pairs in the target block and the comparison block, a first difference factor between the target block and the comparison block is determined; taking the difference between the number of feature points in the target block and the number of matching feature point pairs between the target block and the comparison block as a second difference factor; Based on the first difference factor and the second difference factor, a similarity value between the target block and the comparison block is determined, and both the first difference factor and the second difference factor are negatively correlated with the similarity value.

4. A fingerprint recognition method for intelligent access control according to claim 1, characterized in that: The method for obtaining the matching degree value includes: In the fingerprint image to be identified, the ratio of the number of target blocks whose relative position similarity values ​​are greater than a preset position similarity threshold to the total number of all target blocks is used as a matching degree value.

5. The fingerprint recognition method for intelligent access control according to claim 1 is characterized in that: The fingerprint recognition based on the matching degree value includes: The matching degree values ​​of the fingerprint image to be identified and all the comparison fingerprint images are compared with the preset matching threshold. When there is a matching degree value greater than the preset matching threshold, the fingerprint recognition is judged to be successful, otherwise the fingerprint recognition is judged to be failed.

6. A fingerprint recognition system for intelligent access control, characterized in that: The system comprises: The image acquisition module is used to acquire the fingerprint image to be identified in the intelligent access control fingerprint recognition area; and acquire the comparison fingerprint image pre-entered in the intelligent access control database; A feature matching module is used to divide the fingerprint image to be identified to obtain a target block, divide the comparison fingerprint image to obtain a comparison block; perform feature matching on the target block and the comparison block to obtain a similar block corresponding to the target block; A relative position analysis module is used to determine the neighborhood blocks of each target block in the fingerprint image to be identified, and to determine the neighborhood blocks of each comparison block in the comparison fingerprint image; analyze the similarity of the neighborhood blocks between each target block and the corresponding similar block, and determine the relative position similarity value between the target block and the corresponding similar block; A fingerprint matching module is used to determine the matching degree value between the fingerprint image to be identified and the comparison fingerprint image based on the relative position similarity value between all target blocks and corresponding similar blocks; and perform fingerprint identification based on the matching degree value; The method for obtaining the relative position similarity value comprises: Marking the consistency between the neighborhood blocks of the target block and the neighborhood blocks of the similar blocks corresponding to the target block to obtain a marking sequence corresponding to the target block, and determining the position similarity factor between each target block and the corresponding similar block based on the numerical features in the marking sequence; Based on the corresponding value of each neighboring block of the target block in the tag sequence, the position similarity factor between the target block and the similar block is adjusted to obtain the influence factor of each neighboring block of the target block on the target block; The relative position similarity value between the target block and the corresponding similar block is obtained by combining the position similarity factor between the target block and the corresponding similar block and the influence factor of all the neighboring blocks of all the target blocks on the target block, and the position similarity factor and the influence factor are both positively correlated with the relative position similarity value; The method for obtaining the position similarity factor includes: In the fingerprint image to be identified, the neighborhood block of the target block is used as the reference block; in the fingerprint image to be compared, the neighborhood block of the target block corresponding to the similar block is used as the analysis block; In the analysis block, if there is no analysis block similar to the reference block, all the reference blocks corresponding to the target block are marked as 0, and the marking sequence corresponding to the target block is obtained; If there is an analysis block similar to the reference block, any group of similar analysis blocks and reference blocks are used as the starting block. Starting from the starting block, the similar block of the reference block in the comparison image is judged in a clockwise direction to see whether it is consistent with the analysis block at the corresponding position. If they are consistent, the reference block is marked as 1, otherwise it is marked as 0. The marking sequence corresponding to the target block is obtained. In the tag sequence corresponding to the target block, based on the difference between the number of values ​​1 and the total number of values ​​in the tag sequence, determine a position similarity factor between the target block and the corresponding similar block; The method for obtaining the impact factor includes: The product of the numerical value corresponding to each neighboring block of the target block in the tag sequence and the position similarity factor of the target block is used as the influence factor of the neighboring block of the target block on the target block.

7. A fingerprint recognition device for intelligent access control, characterized in that: The method comprises a processor and a memory, wherein at least one instruction, at least one program, a code set or an instruction set is stored in the memory, and when the at least one instruction, at least one program, a code set or an instruction set is loaded and executed by the processor, the steps of the method as described in any one of claims 1 to 5 are implemented.

Images