Intelligent gate vehicle recognition and identity authentication system and method based on intelligent recognition

Abstract

The present invention relates to the technical field of information identification, and discloses a smart gate vehicle identification and identity authentication system and method based on intelligent identification. First, the initial smart gate vehicle image is subjected to noise reduction processing, image enhancement and license plate coarse positioning to obtain a coarsely positioned smart gate license plate image; secondly, the coarsely positioned smart gate license plate image is positioned based on the grayscale jump method, and then the license plate is segmented; then the license plate features are extracted, and a support vector machine is constructed and trained to obtain a support vector machine classifier, and the license plate features are input into the support vector machine classifier to obtain a vehicle identification result; finally, a smart gate identification database is established, and vehicle identification and identity authentication are performed according to the vehicle identification results. The present invention achieves the purpose of smart gate vehicle identification and identity authentication by processing and analyzing the smart gate vehicle image, and the method is accurate and objective.

Description

Technical Field

[0001] The present invention relates to the technical field of information identification, and in particular to an intelligent gate vehicle identification and identity authentication system and method based on intelligent identification. Background Art

[0002] Chinese patent CN110674527B discloses a privacy protection method, medium, equipment and device based on vehicle owner identity authentication. The method specifically includes: obtaining vehicle identification information and user information associated with the vehicle, generating business information of the current user when the vehicle identification information and the user information associated with the vehicle perform business operations, wherein the business information of the current user includes payment business type, parking lot information, payment amount information and payment number information; calculating the authentication value of the current user and the vehicle according to the business information of the current user, setting an authentication threshold, and when the authentication value of the current user and the vehicle is greater than the authentication threshold, the current user is regarded as the owner of the vehicle and obtains the business operation authority of the vehicle; otherwise, the current user is not the owner of the vehicle, and the owner identity authentication is completed. However, the invention does not further verify the identity, and the result has errors.

[0003] Traditional vehicle identification and identity authentication methods usually use manual identification methods to identify vehicles and personnel, which not only affects the owner's travel experience, but also the identification results are subjective. At the same time, traditional vehicle identification and identity authentication methods do not use high-tech technologies such as neural networks and machine vision, and cannot achieve unmanned and intelligent accurate identification, and work efficiency needs to be improved. Summary of the invention

[0004] In view of the problems in the related art, the present invention provides a smart gate vehicle recognition and identity authentication system and method based on intelligent recognition to overcome the above-mentioned technical problems existing in the existing related art.

[0005] In order to solve the above technical problems, the present invention is achieved through the following technical solutions:

[0006] The present invention is a vehicle identification and identity authentication method for an intelligent gate based on intelligent identification, comprising the following steps:

[0007] S1, the smart gate acquires the image of the vehicle to be detected, obtains the initial smart gate vehicle image, generates the initial smart gate vehicle image matrix, performs image preprocessing and license plate coarse positioning processing based on the initial smart gate vehicle image matrix, and obtains the coarse positioning smart gate license plate image;

[0008] S2, after performing tilt correction on the roughly positioned smart gate license plate image, use the grayscale jump method to locate the license plate to obtain a smart gate license plate positioning image, and then introduce a segmentation threshold to perform license plate segmentation on the smart gate license plate positioning image to obtain a final smart gate license plate image;

[0009] S3, extracting the features of the final intelligent gate license plate image to obtain a license plate feature set, then constructing and training a support vector machine to obtain a support vector machine classifier, performing license plate recognition according to the license plate feature set, obtaining a vehicle recognition result, and realizing vehicle license plate recognition;

[0010] S4. Establish an intelligent gate recognition database, match vehicle license plate information and search for vehicle personnel identity information based on the vehicle recognition result and the intelligent gate recognition database, and complete vehicle recognition and identity authentication.

[0011] The invention obtains the initial intelligent gate vehicle image, first uses the non-local mean algorithm to perform noise reduction processing on the initial intelligent gate vehicle image, then uses histogram equalization to complete image enhancement, and then roughly locates the license plate based on the maximum connected area search algorithm; the non-local mean algorithm realizes noise reduction by calculating the distance weight, and the noise reduction effect is good; the license plate is roughly located first to reduce the amount of subsequent processing calculations and improve the speed of license plate contour detection; secondly, the roughly located intelligent gate license plate image is tilted, located and segmented to obtain the final intelligent gate license plate image; compared with the traditional tilt correction method, the horizontal and vertical directions are corrected during the tilt correction process. It is positive, reduces unnecessary rotation projection, greatly reduces the amount of calculation, and the correction method is simple and accurate; the character segmentation method performs character segmentation according to the projection image, has a wide range of use, and is suitable for complex environments such as rainy days and uneven lighting; then the license plate features are extracted, and a support vector machine classifier is obtained after training to realize vehicle license plate recognition; the license plate features are used to facilitate classifier recognition, and the support vector machine classifier adopts a one-to-many method to achieve multi-class recognition and ensure the minimum classification error rate; finally, an intelligent gate recognition database is established, and vehicle license plate information matching and vehicle and personnel identity information search are performed according to the vehicle recognition results and the intelligent gate recognition database to complete vehicle recognition and identity authentication.

[0012] Preferably, the S1 comprises the following steps:

[0013] S11, monitoring the video at the smart gate to obtain the image of the vehicle to be detected, forming an initial smart gate vehicle image set with each frame of the vehicle image to be detected, selecting the initial smart gate vehicle image, selecting sampling points on the initial smart gate vehicle image, the sampling points being pixels, and performing

[0014] Quantization processing to generate the initial intelligent gate vehicle image matrix A as follows:

[0015] ;

[0016] in, Indicates the horizontal position of the initial intelligent gate vehicle image m , vertical n pixels;

[0017] Based on the initial smart gate vehicle image matrix, the initial smart gate vehicle image is subjected to noise reduction processing using a non-local means algorithm, and then image enhancement is performed using histogram equalization to implement image preprocessing, thereby obtaining a preprocessed smart gate vehicle image matrix. The specific steps are as follows:

[0018] S111, select a size of A search window is formed, and pixels are randomly selected in the search window, which are recorded as the first pixel and the second pixel. The first similar block and the second similar block are respectively established with the first pixel and the second pixel as the center, and the distance weight coefficient of the first similar block and the second similar block is calculated. The calculation formula is as follows:

[0019] ;

[0020] in, represents the distance weight coefficient between the first similar block and the second similar block, Represents the gray value of the first pixel, represents the first pixel, represents the second pixel, e represents a constant;

[0021] Traversing the second similar block, calculating the maximum distance weight coefficient between the pixel point in the second similar block and the first pixel point, assigning the maximum distance weight coefficient to the first pixel point, and performing normalization processing to obtain a normalized result, using the normalized result as the grayscale value of the first pixel point, and completing preliminary noise reduction; repeating S111 until the pixel points in the initial smart gate vehicle image matrix are denoised, and obtaining the denoised smart gate vehicle image matrix;

[0022] S112, plotting a grayscale histogram according to the number of pixels of different grayscale levels in the smart gate vehicle image matrix after denoising to obtain a cumulative probability distribution function, calculating a grayscale transformation function according to the cumulative probability distribution function, and performing image enhancement on the smart gate vehicle image matrix after denoising using the grayscale transformation function to obtain a preprocessed smart gate vehicle image matrix;

[0023] S12, binarize the pre-processed smart gate vehicle image matrix and output the binarized smart gate vehicle image matrix; traverse the binarized smart gate vehicle image matrix, select any pixel point in the binarized smart gate vehicle image matrix, record it as the pixel point to be connected, search the eight neighborhoods of the pixel point to be connected with the pixel point to be connected as the center, when the gray value of the pixel point in the eight neighborhoods is equal to the gray value of the pixel point to be connected, add the corresponding pixel point in the eight neighborhoods to the set of pixel points to be connected, otherwise continue to search other pixel points; until all pixel points in the binarized smart gate vehicle image matrix are searched, a number of sets of pixel points to be connected are obtained, all pixel points in the set of pixel points to be connected are connected to obtain a number of connected areas, mark the largest connected area as the rough positioning area of ​​the license plate, and generate a rough positioning smart gate license plate image.

[0024] The invention first uses a non-local mean algorithm to perform noise reduction on the initial intelligent gate vehicle image, and achieves noise reduction by calculating the distance weight, with good noise reduction effect; then uses histogram equalization to complete image enhancement, and then performs coarse positioning of the license plate based on the maximum connected area search algorithm, thereby reducing the amount of subsequent processing calculations and improving the speed of license plate contour detection.

[0025] Preferably, S2 comprises the following steps:

[0026] S21, using an improved rotational projection method to perform horizontal tilt correction and vertical tilt correction on the roughly positioned smart gate license plate image to obtain an initial smart gate license plate image, the specific steps are as follows:

[0027] S211, setting angle In the range, the step size is , the roughly positioned intelligent gate license plate image is rotated to obtain a first projection value, and the maximum value of the first projection value is obtained as the starting angle, with a step length The roughly positioned intelligent gate license plate image is rotated and projected to obtain a second projection value. When the second projection value is less than or equal to the first projection value, the step size is set to The roughly positioned smart gate license plate image is rotated and projected again until the projection value decreases to obtain a first projection angle, otherwise the rotational projection is continued; the first projection angle is used as the starting angle, and S211 is repeated to obtain a horizontal tilt angle; the roughly positioned smart gate license plate image is rotated at the horizontal tilt angle to obtain a horizontally corrected smart gate license plate image;

[0028] S212, set the vertical projection threshold, at an angle In the range, the step size is , perform vertical projection on the horizontally corrected smart gate license plate image to obtain a third projection value, find the angle corresponding to the maximum value of the third projection value where the third projection value is less than the vertical projection threshold, and record it as the second projection angle; the second projection angle is used as the starting angle, with a step length The horizontally corrected smart gate license plate image is vertically projected again to obtain a fourth projection value; when the fourth projection value is less than or equal to the vertical projection threshold, the step length is The horizontally corrected smart gate license plate image is rotated and projected again until the projection value decreases to obtain a third projection angle, otherwise a vertical projection is performed; the third projection angle is used as the starting angle, and S212 is repeated to obtain a vertical tilt angle; the horizontally corrected smart gate license plate image is tilted and corrected at the vertical tilt angle to obtain an initial smart gate license plate image;

[0029] S22, converting the initial smart gate license plate image into an initial smart gate license plate image matrix, performing jump statistics on each row of the initial smart gate license plate image matrix to obtain a jump array; setting the jump threshold to , select any row in the jump array, record it as the jump record row, and traverse downward from the jump record row. If the number of row jumps is greater than or equal to the jump threshold, the jump record row is recorded as the upper boundary; traverse upward from the jump record row, and when the number of consecutive If the row jump number is greater than or equal to the jump threshold, the jump record row is recorded as the lower boundary;

[0030] The initial smart gate license plate image is vertically projected to obtain a vertical projection image, and the left and right boundary thresholds are set as , the vertical projection image is traversed to the right, when the continuous If the number of row jumps is greater than or equal to the left and right boundary thresholds, the jump record row is recorded as the left boundary; the vertical projection image is traversed to the left, and when the continuous If the row jump number is greater than or equal to the left and right boundary thresholds, the jump record row is recorded as the right boundary;

[0031] In combination with the upper boundary, the lower boundary, the left boundary and the right boundary, the license plate is located on the initial smart gate license plate image to obtain a smart gate license plate location image;

[0032] S23, according to the vertical projection image, counting the average number of grayscale values ​​of 255 in each column of the vertical projection image to obtain the average number of grayscale values, recording one quarter of the average number of grayscale values ​​as the segmentation threshold, marking the positions corresponding to the grayscale values ​​of the pixels less than the segmentation threshold on the vertical projection image to obtain the pre-segmentation positions, calculating the spacing of the pre-segmentation positions to obtain the character width set, and finding the maximum character width and the minimum character width in the character width set;

[0033] A width threshold is set. When the character width in the character width set is less than or equal to the width threshold and greater than or equal to the minimum character width, the pre-segmentation position is retained; otherwise, the character width is reduced by 1 to obtain the first character width, and then the maximum character width is added to obtain the second character width. The character width corresponding to the character width that is less than or equal to the second character width and greater than or equal to the first character width in the character width set is set as the second character width to obtain a new character width set, until the new character width in the new character width set is less than or equal to the width threshold and greater than or equal to the minimum character width to obtain a final character width set; the smart gate license plate positioning image is segmented according to the final character width in the final character width set to obtain a final smart gate license plate image.

[0034] This invention performs tilt correction processing on the roughly positioned smart gate license plate image. Compared with the traditional tilt correction method, the horizontal and vertical directions are corrected separately during the tilt correction process, which reduces unnecessary rotation projection, greatly reduces the amount of calculation, and the correction method is simple and accurate. After license plate positioning and license plate segmentation, the final smart gate license plate image is obtained. The character segmentation method performs character segmentation according to the projection image, which is suitable for complex environments such as rainy days and uneven lighting, and has a wide range of uses.

[0035] Preferably, S3 comprises the following steps:

[0036] S31, using cubic interpolation to normalize the final smart gate license plate image, and then extracting a license plate feature set, the specific steps are as follows:

[0037] S311, setting the pixel coordinates in the final smart gate license plate image to , The gray value of the pixel is , c , d Both represent offset parameters. The grayscale value of the pixels in the final smart gate license plate image is adjusted, and the grayscale value of the adjusted pixels is calculated. The calculation formula is as follows:

[0038] ;

[0039] in, Represents the grayscale value of the pixel after adjustment. express The gray value of the pixel at express The gray value of the pixel at express Gray value of pixel point;

[0040] Generate an adjusted smart gate license plate image according to the adjusted pixel grayscale value;

[0041] S312, counting the number of pixels in each column and the number of pixels in each row in the adjusted smart gate license plate image, and performing normalization processing to obtain a number of character features; and then dividing the adjusted smart gate license plate image into grids, count the number of character pixels and the number of pixels in the grid, calculate the quotient of the number of character pixels and the number of pixels in the grid, and get grid features; combining the character features and the grid features to obtain a license plate feature set;

[0042] S32, obtain a license plate image sample set from the CCDP license plate data set on the Internet, obtain a final license plate image sample set after image preprocessing, license plate coarse positioning processing, tilt correction, license plate positioning and license plate segmentation, extract license plate features of the final license plate image sample set to obtain a license plate feature sample set, divide the license plate feature sample set into a sample training set and a sample test set, construct and train a support vector machine, and obtain a support vector machine classifier. The specific steps are as follows:

[0043] S321, the support vector machine includes a Chinese character classifier, a digital classifier, a digital letter classifier and an alphanumeric classifier, the kernel function is set to use a radial basis function, the radial basis function includes a penalty coefficient and a parameter, the sample training set is input into the support vector machine, and it is continuously iterated until the support vector machine converges to obtain a trained support vector machine;

[0044] S322, input the sample test set into the trained support vector machine, set the accuracy threshold, and when the output result recognition rate is greater than the accuracy threshold, obtain the support vector machine classifier, otherwise adjust the penalty coefficient and parameters until the output result recognition rate is greater than the accuracy threshold;

[0045] S33, inputting the license plate feature set into a support vector machine classifier, outputting a license plate recognition result, and completing vehicle license plate recognition.

[0046] The invention realizes vehicle license plate recognition by extracting license plate features and using the license plate features to train a support vector machine classifier. The license plate features are easy for the classifier to recognize, and the support vector machine classifier adopts a one-to-many method to realize multi-class recognition, ensuring a minimum classification error rate.

[0047] Preferably, S4 comprises the following steps:

[0048] S41, establishing an intelligent gate recognition database, the intelligent gate recognition database includes vehicle license plate information and vehicle personnel identity information, judging whether the license plate recognition result matches the vehicle license plate information, when the license plate recognition result and the vehicle license plate information do not match, the intelligent gate does not allow the vehicle to pass; otherwise, the license plate recognition result and the vehicle license plate information match, and the vehicle recognition is completed;

[0049] S42, after the license plate recognition result matches the vehicle license plate information, the vehicle personnel identity information corresponding to the license plate recognition result is searched in the smart gate recognition database. When the vehicle personnel identity information is searched, the smart gate releases it, otherwise, the identity authentication fails and the identity authentication is completed.

[0050] The present invention also discloses a system of intelligent gate vehicle identification and identity authentication method based on intelligent identification, which specifically includes: an initial intelligent gate vehicle image preprocessing module, an intelligent gate license plate positioning and segmentation module, a vehicle license plate recognition module and a vehicle identification and identity authentication module;

[0051] The initial intelligent gate vehicle image preprocessing module is used for performing noise reduction and histogram equalization on the initial intelligent gate vehicle image;

[0052] The smart gate license plate positioning and segmentation module is used to use the grayscale jump method to locate the license plate, and then introduce the segmentation threshold to segment the license plate;

[0053] The vehicle license plate recognition module is used to construct and train a support vector machine classifier to recognize the license plate;

[0054] The vehicle identification and identity authentication module is used to match vehicle license plate information and search for vehicle personnel identity information based on a database.

[0055] The present invention has the following beneficial effects:

[0056] 1. The invention first uses the non-local mean algorithm to perform noise reduction on the initial intelligent gate vehicle image, and realizes noise reduction by calculating the distance weight, with good noise reduction effect; then uses histogram equalization to complete image enhancement, and performs rough positioning of the license plate based on the maximum connected area search algorithm, thereby reducing the amount of subsequent processing calculations and improving the speed of license plate contour detection.

[0057] 2. The invention performs tilt correction processing on the roughly positioned intelligent gate license plate image. Compared with the traditional tilt correction method, the horizontal and vertical directions are corrected separately during the tilt correction process, which reduces unnecessary rotation projection and greatly reduces the amount of calculation. The correction method is simple and accurate; then the license plate positioning and license plate segmentation are performed. The character segmentation method performs character segmentation according to the projected image, which is suitable for complex environments such as rainy days and uneven lighting, and has a wide range of uses.

[0058] 3. The invention realizes vehicle license plate recognition by extracting license plate features and using the license plate features to train a support vector machine classifier. The license plate features are easy for the classifier to recognize, and the support vector machine classifier adopts a one-to-many method to realize multi-class recognition, ensuring a minimum classification error rate.

[0059] Of course, any product implementing the present invention does not necessarily need to achieve all of the advantages described above at the same time. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] In order to more clearly illustrate the technical solutions of the embodiments of the invention, the drawings required for describing the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the invention. For ordinary technicians in this field, they can also obtain drawings based on these drawings without paying creative work.

[0061] Figure 1 A schematic diagram of the process of vehicle identification and identity authentication at a smart gate by the smart gate vehicle identification and identity authentication system based on intelligent identification provided by the present invention. DETAILED DESCRIPTION

[0062] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0063] In the description of the present invention, it is necessary to understand that the terms "opening", "upper", "lower", "top", "middle", "inside" and the like indicating orientation or positional relationship are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the components or elements referred to must have a specific orientation, be constructed and operate in a specific orientation, and therefore cannot be understood as limiting the invention. Example

[0064] Please refer to Figure 1 The vehicle identification and identity authentication method of the intelligent gate based on intelligent identification of the present invention comprises the following steps:

[0065] S1, the smart gate acquires the image of the vehicle to be detected, obtains the initial smart gate vehicle image, generates the initial smart gate vehicle image matrix, performs image preprocessing and license plate coarse positioning processing based on the initial smart gate vehicle image matrix, and obtains the coarse positioning smart gate license plate image;

[0066] The S1 comprises the following steps:

[0067] S11, monitoring the video at the smart gate to obtain the image of the vehicle to be detected, forming an initial smart gate vehicle image set with each frame of the vehicle image to be detected, selecting the initial smart gate vehicle image, selecting sampling points on the initial smart gate vehicle image, the sampling points being pixels, and performing

[0068] Quantization processing to generate the initial intelligent gate vehicle image matrix A as follows:

[0069] ;

[0070] in, Indicates the horizontal position of the initial intelligent gate vehicle image m , vertical n pixels;

[0071] Based on the initial smart gate vehicle image matrix, the initial smart gate vehicle image is subjected to noise reduction processing using a non-local means algorithm, and then image enhancement is performed using histogram equalization to implement image preprocessing, thereby obtaining a preprocessed smart gate vehicle image matrix. The specific steps are as follows:

[0072] S111, select a size of A search window is formed, and pixels are randomly selected in the search window, which are recorded as the first pixel and the second pixel. The first similar block and the second similar block are respectively established with the first pixel and the second pixel as the center, and the distance weight coefficient of the first similar block and the second similar block is calculated. The calculation formula is as follows:

[0073] ;

[0074] in, represents the distance weight coefficient between the first similar block and the second similar block, Represents the gray value of the first pixel, represents the first pixel, represents the second pixel, e represents a constant;

[0075] Traversing the second similar block, calculating the maximum distance weight coefficient between the pixel point in the second similar block and the first pixel point, assigning the maximum distance weight coefficient to the first pixel point, and performing normalization processing to obtain a normalized result, using the normalized result as the grayscale value of the first pixel point, and completing preliminary noise reduction; repeating S111 until the pixel points in the initial smart gate vehicle image matrix are denoised, and obtaining the denoised smart gate vehicle image matrix;

[0076] S112, plotting a grayscale histogram according to the number of pixels of different grayscale levels in the smart gate vehicle image matrix after denoising to obtain a cumulative probability distribution function, calculating a grayscale transformation function according to the cumulative probability distribution function, and performing image enhancement on the smart gate vehicle image matrix after denoising using the grayscale transformation function to obtain a preprocessed smart gate vehicle image matrix;

[0077] S12, binarize the pre-processed intelligent gate vehicle image matrix and output the binary intelligent gate vehicle image matrix; traverse the binary intelligent gate vehicle image matrix, select any pixel point in the binary intelligent gate vehicle image matrix, record it as the pixel point to be connected, search the eight neighborhoods of the pixel point to be connected with the pixel point to be connected as the center, when the gray value of the pixel point in the eight neighborhoods is equal to the gray value of the pixel point to be connected, add the corresponding pixel point in the eight neighborhoods to the set of pixel points to be connected, otherwise continue to search other pixel points; until all pixel points in the binary intelligent gate vehicle image matrix are searched, a number of sets of pixel points to be connected are obtained, all pixel points in the set of pixel points to be connected are connected, a number of connected areas are obtained, the largest connected area is marked as the rough positioning area of ​​the license plate, and a rough positioning intelligent gate license plate image is generated;

[0078] S2, after performing tilt correction on the roughly positioned smart gate license plate image, use the grayscale jump method to locate the license plate to obtain a smart gate license plate positioning image, and then introduce a segmentation threshold to perform license plate segmentation on the smart gate license plate positioning image to obtain a final smart gate license plate image;

[0079] The S2 comprises the following steps:

[0080] S21, using an improved rotational projection method to perform horizontal tilt correction and vertical tilt correction on the roughly positioned smart gate license plate image to obtain an initial smart gate license plate image, the specific steps are as follows:

[0081] S211, setting angle In the range, the step size is , the roughly positioned intelligent gate license plate image is rotated to obtain a first projection value, and the maximum value of the first projection value is obtained as the starting angle, with a step length The roughly positioned intelligent gate license plate image is rotated and projected to obtain a second projection value. When the second projection value is less than or equal to the first projection value, the step size is set to The roughly positioned smart gate license plate image is rotated and projected again until the projection value decreases to obtain a first projection angle, otherwise the rotational projection is continued; the first projection angle is used as the starting angle, and S211 is repeated to obtain a horizontal tilt angle; the roughly positioned smart gate license plate image is rotated at the horizontal tilt angle to obtain a horizontally corrected smart gate license plate image;

[0082] S212, set the vertical projection threshold, at an angle In the range, the step size is , perform vertical projection on the horizontally corrected smart gate license plate image to obtain a third projection value, find the angle corresponding to the maximum value of the third projection value where the third projection value is less than the vertical projection threshold, and record it as the second projection angle; the second projection angle is used as the starting angle, with a step length The horizontally corrected smart gate license plate image is vertically projected again to obtain a fourth projection value; when the fourth projection value is less than or equal to the vertical projection threshold, the step length is The horizontally corrected smart gate license plate image is rotated and projected again until the projection value decreases to obtain a third projection angle, otherwise a vertical projection is performed; the third projection angle is used as the starting angle, and S212 is repeated to obtain a vertical tilt angle; the horizontally corrected smart gate license plate image is tilted and corrected at the vertical tilt angle to obtain an initial smart gate license plate image;

[0083] S22, converting the initial smart gate license plate image into an initial smart gate license plate image matrix, performing jump statistics on each row of the initial smart gate license plate image matrix to obtain a jump array; setting the jump threshold to , select any row in the jump array, record it as the jump record row, and traverse downward from the jump record row. If the number of row jumps is greater than or equal to the jump threshold, the jump record row is recorded as the upper boundary; traverse upward from the jump record row, and when the number of consecutive If the row jump number is greater than or equal to the jump threshold, the jump record row is recorded as the lower boundary;

[0084] The initial smart gate license plate image is vertically projected to obtain a vertical projection image, and the left and right boundary thresholds are set as , the vertical projection image is traversed to the right, when the continuous If the number of row jumps is greater than or equal to the left and right boundary thresholds, the jump record row is recorded as the left boundary; the vertical projection image is traversed to the left, and when the continuous If the row jump number is greater than or equal to the left and right boundary thresholds, the jump record row is recorded as the right boundary;

[0085] In combination with the upper boundary, the lower boundary, the left boundary and the right boundary, the license plate is located on the initial smart gate license plate image to obtain a smart gate license plate location image;

[0086] S23, according to the vertical projection image, counting the average number of grayscale values ​​of 255 in each column of the vertical projection image to obtain the average number of grayscale values, recording one quarter of the average number of grayscale values ​​as the segmentation threshold, marking the positions corresponding to the grayscale values ​​of the pixels less than the segmentation threshold on the vertical projection image to obtain the pre-segmentation positions, calculating the spacing of the pre-segmentation positions to obtain the character width set, and finding the maximum character width and the minimum character width in the character width set;

[0087] A width threshold is set. When the character width in the character width set is less than or equal to the width threshold and greater than or equal to the minimum character width, the pre-segmentation position is retained; otherwise, the character width is reduced by 1 to obtain the first character width, and the maximum character width is added to obtain the second character width. The character widths corresponding to the character widths less than or equal to the second character width and greater than or equal to the first character width in the character width set are set as the second character width to obtain a new character width set, until the new character width in the new character width set is less than or equal to the width threshold and greater than or equal to the minimum character width, to obtain a final character width set; the smart gate license plate positioning image is segmented according to the final character width in the final character width set to obtain a final smart gate license plate image;

[0088] S3, extracting the features of the final intelligent gate license plate image to obtain a license plate feature set, then constructing and training a support vector machine to obtain a support vector machine classifier, performing license plate recognition according to the license plate feature set, obtaining a vehicle recognition result, and realizing vehicle license plate recognition;

[0089] The S3 comprises the following steps:

[0090] S31, using cubic interpolation to normalize the final smart gate license plate image, and then extracting a license plate feature set, the specific steps are as follows:

[0091] S311, setting the pixel coordinates in the final smart gate license plate image to , The gray value of the pixel is , c , d Both represent offset parameters. The grayscale value of the pixels in the final smart gate license plate image is adjusted, and the grayscale value of the adjusted pixels is calculated. The calculation formula is as follows:

[0092] ;

[0093] in, Represents the grayscale value of the pixel after adjustment. express The gray value of the pixel at express The gray value of the pixel at express Gray value of pixel point;

[0094] Generate an adjusted smart gate license plate image according to the adjusted pixel grayscale value;

[0095] S312, counting the number of pixels in each column and the number of pixels in each row in the adjusted smart gate license plate image, and performing normalization processing to obtain a number of character features; and then dividing the adjusted smart gate license plate image into grids, count the number of character pixels and the number of pixels in the grid, calculate the quotient of the number of character pixels and the number of pixels in the grid, and get grid features; combining the character features and the grid features to obtain a license plate feature set;

[0096] S32, obtain a license plate image sample set from the CCDP license plate data set on the Internet, obtain a final license plate image sample set after image preprocessing, license plate coarse positioning processing, tilt correction, license plate positioning and license plate segmentation, extract license plate features of the final license plate image sample set to obtain a license plate feature sample set, divide the license plate feature sample set into a sample training set and a sample test set, construct and train a support vector machine, and obtain a support vector machine classifier. The specific steps are as follows:

[0097] S321, the support vector machine includes a Chinese character classifier, a digital classifier, a digital letter classifier and an alphanumeric classifier, the kernel function is set to use a radial basis function, the radial basis function includes a penalty coefficient and a parameter, the sample training set is input into the support vector machine, and it is continuously iterated until the support vector machine converges to obtain a trained support vector machine;

[0098] S322, input the sample test set into the trained support vector machine, set the accuracy threshold, and when the output result recognition rate is greater than the accuracy threshold, obtain the support vector machine classifier, otherwise adjust the penalty coefficient and parameters until the output result recognition rate is greater than the accuracy threshold;

[0099] S33, inputting the license plate feature set into a support vector machine classifier, outputting a license plate recognition result, and completing vehicle license plate recognition;

[0100] S4, establish an intelligent gate recognition database, match the vehicle license plate information and search the vehicle personnel identity information according to the vehicle recognition result and the intelligent gate recognition database, and complete vehicle recognition and identity authentication;

[0101] The S4 comprises the following steps:

[0102] S41, establishing an intelligent gate recognition database, the intelligent gate recognition database includes vehicle license plate information and vehicle personnel identity information, judging whether the license plate recognition result matches the vehicle license plate information, when the license plate recognition result and the vehicle license plate information do not match, the intelligent gate does not allow the vehicle to pass; otherwise, the license plate recognition result and the vehicle license plate information match, and the vehicle recognition is completed;

[0103] S42, after the license plate recognition result matches the vehicle license plate information, the vehicle personnel identity information corresponding to the license plate recognition result is searched in the smart gate recognition database. When the vehicle personnel identity information is searched, the smart gate releases it, otherwise, the identity authentication fails and the identity authentication is completed. Example

[0104] The present invention also discloses a system of intelligent gate vehicle identification and identity authentication method based on intelligent identification, which specifically includes: an initial intelligent gate vehicle image preprocessing module, an intelligent gate license plate positioning and segmentation module, a vehicle license plate recognition module and a vehicle identification and identity authentication module;

[0105] The initial intelligent gate vehicle image preprocessing module is used for performing noise reduction and histogram equalization on the initial intelligent gate vehicle image;

[0106] The smart gate license plate positioning and segmentation module is used to use the grayscale jump method to locate the license plate, and then introduce the segmentation threshold to segment the license plate;

[0107] The vehicle license plate recognition module is used to construct and train a support vector machine classifier to recognize the license plate;

[0108] The vehicle identification and identity authentication module is used to match vehicle license plate information and search for vehicle personnel identity information based on a database.

[0109] In the description of this specification, the description with reference to the terms "one embodiment", "example", "specific example", etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0110] The preferred embodiments of the invention disclosed above are only used to help explain the invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to the specific implementation methods described. Obviously, many modifications and changes can be made according to the content of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the invention, so that those skilled in the art can understand and use the invention well.

Claims

1. A vehicle identification and identity authentication method for an intelligent gate based on intelligent identification, characterized in that: The steps include: S1, the smart gate acquires the image of the vehicle to be detected, obtains the initial smart gate vehicle image, generates the initial smart gate vehicle image matrix, performs image preprocessing and license plate coarse positioning processing based on the initial smart gate vehicle image matrix, and obtains the coarse positioning smart gate license plate image; The S1 comprises the following steps: S11 includes the following steps: S111, selecting a search window on the initial smart gate vehicle image matrix, selecting a similar block in the search window, recording the central pixel point of the similar block as the first pixel point, and calculating the distance weight coefficient of the similar block; assigning the maximum distance weight coefficient to the first pixel point, and performing normalization processing to complete preliminary noise reduction, until the pixel points in the initial smart gate vehicle image matrix are completely denoised, and obtaining the denoised smart gate vehicle image matrix; S112, plotting a grayscale histogram according to the number of pixels of different grayscale levels in the smart gate vehicle image matrix after denoising to obtain a cumulative probability distribution function, calculating a grayscale transformation function according to the cumulative probability distribution function, and performing image enhancement on the smart gate vehicle image matrix after denoising using the grayscale transformation function to obtain a preprocessed smart gate vehicle image matrix; S12, binarizing the pre-processed smart gate vehicle image matrix, outputting the binarized smart gate vehicle image matrix, searching the eight neighborhoods of the pixel points in the binarized smart gate vehicle image matrix, obtaining a number of connected areas, marking the largest connected area as the license plate coarse positioning area, completing the license plate coarse positioning processing, and generating a coarsely positioned smart gate license plate image; S2, after performing tilt correction on the roughly positioned smart gate license plate image, use the grayscale jump method to locate the license plate to obtain a smart gate license plate positioning image, and then introduce a segmentation threshold to perform license plate segmentation on the smart gate license plate positioning image to obtain a final smart gate license plate image; S3, extracting the features of the final intelligent gate license plate image to obtain a license plate feature set, then constructing and training a support vector machine to obtain a support vector machine classifier, performing license plate recognition according to the license plate feature set, obtaining a vehicle recognition result, and realizing vehicle license plate recognition; S4. Establish an intelligent gate recognition database, match vehicle license plate information and search for vehicle personnel identity information based on the vehicle recognition result and the intelligent gate recognition database, and complete vehicle recognition and identity authentication.

2. The intelligent gate vehicle recognition and identity authentication method based on intelligent recognition according to claim 1 is characterized in that: The S2 comprises the following steps: S21, using an improved rotational projection method to perform horizontal tilt correction and vertical tilt correction on the roughly positioned smart gate license plate image to obtain an initial smart gate license plate image; S22, performing jump statistics on the initial smart gate license plate image to obtain a jump array; Setting a transition threshold, comparing the number of transition rows in the transition array with the transition threshold, obtaining an upper boundary, a lower boundary, a left boundary and a right boundary, performing license plate positioning on the initial smart gate license plate image, and obtaining a smart gate license plate positioning image; S23, vertically projecting the smart gate license plate positioning image to obtain a vertical projection image and a character width set; A width threshold is set, and the relationship between the width threshold and the maximum and minimum character widths is compared, and the character width set is updated to obtain a final character width set; the smart gate license plate positioning image is segmented according to the final character width set to obtain a final smart gate license plate image.

3. The intelligent gate vehicle recognition and identity authentication method based on intelligent recognition according to claim 2 is characterized in that: The S21 comprises the following steps: S211, rotating the roughly positioned smart gate license plate image to obtain a horizontal tilt angle, rotating the roughly positioned smart gate license plate image at the horizontal tilt angle to obtain a horizontally corrected smart gate license plate image; S212, vertically projecting the horizontally corrected smart gate license plate image to obtain a vertical tilt angle, and performing tilt correction on the horizontally corrected smart gate license plate image at the vertical tilt angle to obtain an initial smart gate license plate image.

4. The intelligent gate vehicle recognition and identity authentication method based on intelligent recognition according to claim 3 is characterized in that: The S3 comprises the following steps: S31, using cubic interpolation to normalize the final smart gate license plate image, and then extracting a license plate feature set; S32, obtaining a license plate image sample set from the CCDP license plate data set on the Internet, obtaining a final license plate image sample set after image preprocessing, license plate coarse positioning processing, tilt correction, license plate positioning and license plate segmentation, extracting license plate features of the final license plate image sample set to obtain a license plate feature sample set, dividing the license plate feature sample set into a sample training set and a sample test set, constructing and training a support vector machine, and obtaining a support vector machine classifier; S33, inputting the license plate feature set into a support vector machine classifier, outputting a license plate recognition result, and completing vehicle license plate recognition.

5. The intelligent gate vehicle recognition and identity authentication method based on intelligent recognition according to claim 4 is characterized in that: The S31 comprises the following steps: S311, using cubic interpolation to adjust the grayscale values ​​of the pixels in the final smart gate license plate image, calculating the adjusted grayscale values ​​of the pixels, and generating the adjusted smart gate license plate image; S312, counting the number of pixels in each column and the number of pixels in each row in the adjusted smart gate license plate image, and performing normalization processing to obtain a number of character features; and then dividing the adjusted smart gate license plate image into grids, count the number of character pixels and the number of pixels in the grid, calculate the quotient of the number of character pixels and the number of pixels in the grid, and get grid features; combining the character features and the grid features to obtain a license plate feature set.

6. The intelligent gate vehicle recognition and identity authentication method based on intelligent recognition according to claim 5 is characterized in that: The construction and training of a support vector machine to obtain a support vector machine classifier comprises the following steps: The support vector machine includes a Chinese character classifier, a digital classifier, a digital letter classifier and an alphanumeric classifier. The kernel function is set to use a radial basis function, and the radial basis function includes a penalty coefficient and a parameter. The sample training set is input into the support vector machine, and it is continuously iterated until the support vector machine converges to obtain a trained support vector machine. The sample test set is input into the trained support vector machine, and an accuracy threshold is set. When the output result recognition rate is greater than the accuracy threshold, a support vector machine classifier is obtained. Otherwise, the penalty coefficient and parameters are adjusted until the output result recognition rate is greater than the accuracy threshold.

7. The intelligent gate vehicle recognition and identity authentication method based on intelligent recognition according to claim 6 is characterized in that: The S4 comprises the following steps: S41, establishing an intelligent gate recognition database, matching the vehicle license plate information according to the vehicle recognition result and the intelligent gate recognition database, and completing vehicle recognition; S42, after the vehicle is identified, the vehicle and personnel identity information corresponding to the license plate recognition result is searched in the smart gate recognition database. When the vehicle and personnel identity information is found, the smart gate allows the vehicle to pass. Otherwise, the identity authentication fails and the identity authentication is completed.

8. A system for implementing the intelligent gate vehicle recognition and identity authentication method based on intelligent identification as claimed in any one of claims 1 to 7, characterized in that: Specifically include: Initial intelligent gate vehicle image preprocessing module, intelligent gate license plate positioning and segmentation module, vehicle license plate recognition module and vehicle recognition and identity authentication module; The initial intelligent gate vehicle image preprocessing module is used for performing noise reduction and histogram equalization on the initial intelligent gate vehicle image; The smart gate license plate positioning and segmentation module is used to use the grayscale jump method to locate the license plate, and then introduce the segmentation threshold to segment the license plate; The vehicle license plate recognition module is used to construct and train a support vector machine classifier to recognize the license plate; The vehicle identification and identity authentication module is used to match vehicle license plate information and search for vehicle personnel identity information based on a database.

Images