A data distribution management system and method in large-scale scenarios

Abstract

The present invention discloses a data distribution management system and method in a large-scale scenario, which belongs to the technical field of data distribution management systems, and includes a data acquisition module, a data analysis and processing module, a feature comparison and judgment module, a verification notification module, and a database update module. The data acquisition module is used to collect data, the data analysis and processing module is used to analyze and process the collected data, the feature comparison and judgment module is used to compare and match the data features of the collected information with the corresponding feature information in the original database, the verification notification module is used to notify the staff in real time of the comparison and matching results of the feature information and to perform on-site verification of the information mismatch, the database update module is used to store new data and complete the database update, and when the image information collected in real time shows abnormal matching of handheld objects, the system will remind the staff to perform manual verification, and the database update is performed after the verification is passed.

Description

Technical Field

[0001] The present invention relates to the technical field of data distribution management systems, and in particular to a data distribution management system and method in a large-scale scenario. Background Art

[0002] With the economic growth and social development, people often organize large-scale activities for entertainment, publicity, learning, health testing, etc. However, some large-scale activities have some restrictions on the participants. Due to the large number of participants, there will be the phenomenon of others pretending to participate. It is very important to quickly and effectively detect the identity of the participants.

[0003] The existing data distribution management system mainly uses the method of scanning the ID card to obtain the QR code with identity information for identity authentication, which makes some people use other people's ID cards or QR codes for detection, resulting in some major security risks;

[0004] Therefore, a data distribution management system and method in large-scale scenarios is needed to solve the above problems. Summary of the invention

[0005] The purpose of the present invention is to provide a data distribution management system in a large-scale scenario to solve the problems raised in the above background technology.

[0006] In order to achieve the above-mentioned object, the present invention provides the following technical solutions: a data distribution management system in a large-scale scenario, the data distribution management system includes a data acquisition module, a data analysis and processing module, a feature comparison and judgment module, a verification notification module, and a database update module;

[0007] Enables accurate identity recognition in large-scale scenarios;

[0008] The data acquisition module is used to transmit the data collected in large-scale scenarios to the data analysis and processing module;

[0009] The data analysis and processing module is used to analyze and process the collected data;

[0010] The feature comparison and judgment module is used to compare and match the feature information of the collected information with the corresponding feature information in the original database;

[0011] The verification notification module is used to notify the staff of the comparison and matching results of the feature information in real time, and to perform on-site verification of mismatches;

[0012] The database update module stores the collected data into the original database;

[0013] The output end of the data acquisition module is connected to the input end of the data analysis and processing module; the output end of the data analysis and processing module is connected to the input end of the feature comparison and judgment module; the output end of the feature comparison module is connected to the input end of the verification notification module, and the output end of the verification notification module is connected to the input end of the database update module.

[0014] According to the above technical solution, the data acquisition module collects data through an image input device and an image scanning device; the image input device and the image scanning device refer to scanning devices with cameras such as mobile phones, and transmit the collected data to the data analysis and processing module.

[0015] According to the above technical solution, the data analysis and processing module includes an information extraction unit, a feature extraction unit, and a data storage unit;

[0016] The information extraction unit is used to extract useful personal information from the collected data, the personal information including name, gender, home address, ID number, headshot, telephone number, QR code information, so that the personal identity information is more complete, and transmit the extracted personal information to the data storage unit;

[0017] The feature extraction unit is used to extract features from the collected data and transmit the extracted feature information to the data storage unit;

[0018] The data storage unit is used to store the extracted personal information and characteristic information;

[0019] The output end of the information extraction unit is connected to the input end of the feature extraction unit; the output end of the feature extraction unit is connected to the input end of the data storage unit.

[0020] According to the above technical solution, the feature comparison and judgment module includes an information matching unit, a handheld object matching unit, a hand feature matching unit, and an angle matching unit;

[0021] The information matching unit matches the collected personal information with the personal information in the original database; the original database refers to the database before the collected data is stored in the database;

[0022] The handheld object matching unit matches the handheld object information in the collected image with the corresponding handheld object information in the original database; the handheld object refers to an ID card and a mobile phone;

[0023] The hand feature matching unit is used to match the hand features of the collected image with the corresponding feature information in the original database; the corresponding feature information refers to the feature information when the user information in the collected image matches the personal information in the original database;

[0024] The angle matching unit is used to match the angle of the handheld object with the angle of the corresponding handheld object in the original database; the angle of the handheld object refers to the angle at which the handheld object is held in the captured image;

[0025] The feature comparison and judgment module performs four feature matches, so that the judgment result of the feature comparison and judgment module is more accurate and more convincing;

[0026] The output end of the personal information matching unit is connected to the input ends of the handheld object matching unit, the hand feature matching unit, and the angle matching unit.

[0027] According to the above technical solution, the verification notification module is used to notify the staff of the result of feature matching in real time and perform on-site verification of mismatches, so that the staff can go and handle it as soon as possible.

[0028] According to the above technical solution, the database update module completes the database update by saving the collected correct information to the corresponding position of the original database; the corresponding position of the original database refers to storing the feature information in the collected image in the storage unit in the original database that matches the personal user information with the personal information box in the collected image.

[0029] A data distribution management method in a large-scale scenario, the data distribution management method comprising the following steps:

[0030] S1: Use image input equipment and image scanning equipment to collect image information and transmit it to the data analysis and processing module in real time;

[0031] S2: extracting personal information from the collected image using the information extraction unit; extracting features from the collected image information using the feature extraction unit; and storing the extracted personal information and feature information using the data storage unit;

[0032] S3: using the information matching unit to match the personal information in the original database, and after the personal information is successfully matched, matching is performed with the corresponding feature information in the original database through the handheld object matching unit, the hand feature matching unit, and the angle matching unit;

[0033] S4: Real-time notification based on personal information and feature matching results in S3, and reminder for manual verification in case of mismatch;

[0034] S5: Intelligently update the database according to the verification results in S4.

[0035] 9. According to the above technical solution, the extraction of personal information in the collected image in step S2 is divided into two types, one is to automatically identify and read the personal information in the QR code through the image input device and the image scanning device, and the other is to use the dlib library to identify the face of the picture after preprocessing the picture, find the eye feature points and calculate the inclination angle of the eye, rotate the photo and extract the head portrait and convert the image into a binary image, and then use the pytesseract library to identify the text information on the picture, and extract the personal information through the print function; the feature extraction includes handheld object feature extraction, hand feature extraction, and angle extraction; the handheld object feature extraction and hand feature extraction are extracted using a convolutional neural network; the binary image is a special grayscale image, that is, the grayscale value of any pixel is 0 or 255, representing black and white respectively; the dlib library, pytesseract library, print function, and convolutional neural network are all known technologies and are not specifically described here; the image preprocessing text image is separated and handed over to the recognition module for recognition, in order to eliminate useless information in the image, retain useful information and simplify data to the maximum extent;

[0036] The angle extraction refers to identifying and extracting the handheld object in the captured image, and the camera of the image input device and the image scanning device is parallel to the extract to collect the image, and the collected picture of the extract is processed and stored as a standard picture in the standard object storage unit of the database, and the vertex of the upper left corner of the standard picture is used as the starting point to assign four vertices ABCD in a clockwise order, and a coordinate system is established with point D as the origin, DC as the positive direction of the x-axis, and DA as the positive direction of the y-axis, and the intersection of AC and BD is connected as the center point M, the coordinates of point M are represented by (a, b), and the coordinates of point A are represented by (x A ,y A ), the coordinates of point C are represented by (x C ,y C ) indicates that its center point The center point M is used as the standard reference point, and the lengths of DC and DA are measured to be L1 and L2 respectively;

[0037] The extract of the handheld object in the collected image is numbered according to the corresponding position in the standard picture of the rectangular coordinate system. A corresponds to E, B corresponds to F, C corresponds to G, and D corresponds to H. A rectangular coordinate system is established with H as the origin, the horizontal direction as the x-axis, and the position perpendicular to the horizontal direction as the y-axis. The lengths l1 and l2 of HE and HG are measured, and l1 and l2 are compared with L1 and L2. If they are not equal, the extracted image is enlarged in the same proportion so that l1 = L1 and l2 = L2. EG and FH are connected and the intersection is taken as the center point Q. The coordinates of point Q are represented by (m, n), and the coordinates of point E are represented by (x E ,y E ), the coordinates of point G are represented by (xG ,y G ) indicates that its center point Put the M and Q coordinates into the same rectangular coordinate system with the origin at O, then ∠MOQ is the extracted tilt angle of the handheld object.

[0038] According to the above technical solution, the feature matching in step S3 is based on area matching, and the angle matching is used to compare the angle information in the original database with the angle information extracted from the captured image. Put into the set K = {θ1,θ2,……,θ n-1}, and average the data in the K sets to get the optimal angle δ. The angle matching rate is: Said ε is the minimum maximum error of the angle matching rate;

[0039] when Indicates that the angle matching is successful;

[0040] When other situations occur, it means that the angle matching fails.

[0041] According to the above technical solution, step S4 notifies the matching result of S3 in real time. If a mismatch occurs, the staff is reminded to perform manual verification. After the manual verification is passed, step S5 updates the database, for example, by voice reminder, pop-up prompt, or flashing warning light.

[0042] Through the above technical solution, accurate and rapid identity recognition can be performed in large-scale scenarios, and impersonation by others can be effectively identified, effectively reducing safety hazards and avoiding dangerous accidents.

[0043] Compared with the prior art, the beneficial effects achieved by the present invention are:

[0044] 1. This invention is equipped with an information matching unit, a handheld object matching unit, a hand feature matching unit, and an angle matching unit to match multiple feature information of the collected images, so that the matching results are more accurate and more convincing, and the identity recognition is more precise, which can avoid some major losses and safety hazards;

[0045] 2. This invention sets up a verification notification module, which can notify the on-site staff of the image feature matching results to take timely measures, and timely remind and conduct on-site verification of mismatches, so as to avoid misjudgment;

[0046] 3. This invention sets up a database update module to update the database in real time, which can make the feature information in the database richer and more detailed, and make feature matching more accurate. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0048] Figure 1 It is a schematic diagram of the module composition of a data distribution management system in a large-scale scenario of the present invention;

[0049] Figure 2 It is a schematic diagram of the steps of a data distribution management method in a large-scale scenario of the present invention;

[0050] Figure 3 It is a schematic diagram of the connection structure of a data distribution management system in a large-scale scenario of the present invention. DETAILED DESCRIPTION

[0051] 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.

[0052] like Figure 1 to Figure 3 As shown, the present invention provides the following technical solution: a data distribution management system in a large-scale scenario, the data distribution management system includes a data acquisition module, a data analysis and processing module, a feature comparison and judgment module, a verification notification module, and a database update module;

[0053] Enables accurate identity recognition in large-scale scenarios;

[0054] The data acquisition module is used to transmit the data collected in large-scale scenarios to the data analysis and processing module;

[0055] The data analysis and processing module is used to analyze and process the collected data;

[0056] The feature comparison and judgment module is used to compare and match the feature information of the collected information with the corresponding feature information in the original database;

[0057] The verification notification module is used to notify the staff of the comparison and matching results of the feature information in real time, and to perform on-site verification of mismatches;

[0058] The database update module stores the collected data into the original database;

[0059] The output end of the data acquisition module is connected to the input end of the data analysis and processing module; the output end of the data analysis and processing module is connected to the input end of the feature comparison and judgment module; the output end of the feature comparison module is connected to the input end of the verification notification module, and the output end of the verification notification module is connected to the input end of the database update module.

[0060] The data acquisition module acquires data through an image input device and an image scanning device; the image input device and the image scanning device refer to scanning devices with cameras such as mobile phones, and transmits the acquired data to the data analysis and processing module.

[0061] The data analysis and processing module includes an information extraction unit, a feature extraction unit, and a data storage unit;

[0062] The information extraction unit is used to extract useful personal information from the collected data, the personal information including name, gender, home address, ID number, headshot, telephone number, QR code information, so that the personal identity information is more complete, and transmit the extracted personal information to the data storage unit;

[0063] The feature extraction unit is used to extract features from the collected data and transmit the extracted feature information to the data storage unit;

[0064] The data storage unit is used to store the extracted personal information and feature information, for example, extracting personal information and feature information from the collected data, and storing the extracted feature information in the personal information storage unit;

[0065] The output end of the information extraction unit is connected to the input end of the feature extraction unit; the output end of the feature extraction unit is connected to the input end of the data storage unit.

[0066] The feature comparison and judgment module includes an information matching unit, a handheld object matching unit, a hand feature matching unit, and an angle matching unit;

[0067] The information matching unit is used to match the collected personal information with the personal information in the original database; the original database refers to the database before the collected data is stored in the database. For example, after the collected image is compared with the features and the staff is notified, the system automatically stores the data in the original database under the personal information storage unit with neutral matching;

[0068] The handheld object matching unit matches the handheld object information in the collected image with the corresponding handheld object information in the original database; the handheld object refers to an ID card or a mobile phone. For example, in a large-scale health detection activity, everyone holds a mobile phone to scan a QR code or holds an ID card for identity recognition;

[0069] The hand feature matching unit is used to match the hand features of the collected image with the corresponding feature information in the original database; the corresponding feature information refers to the feature information when the user information in the collected image matches the personal information in the original database;

[0070] The angle matching unit is used to match the angle of the handheld object with the angle of the corresponding handheld object in the original database; the angle of the handheld object refers to the angle at which the handheld object is held in the captured image;

[0071] The feature comparison and judgment module performs four feature matches, so that the judgment result of the feature comparison and judgment module is more accurate and more convincing;

[0072] The output end of the personal information matching unit is connected to the input ends of the handheld object matching unit, the hand feature matching unit, and the angle matching unit.

[0073] The verification notification module is used to notify the staff of the result of feature matching in real time and perform on-site verification of mismatches, so that the staff can go and handle it as soon as possible.

[0074] The database update module completes the database update by saving the correct information collected to the corresponding position of the original database; the corresponding position of the original database refers to storing the feature information in the collected image in the storage unit in the original database that matches the personal user information with the personal information box in the collected image.

[0075] A data distribution management method in a large-scale scenario, the data distribution management method comprising the following steps:

[0076] S1: Use image input equipment and image scanning equipment to collect image information and transmit it to the data analysis and processing module in real time;

[0077] S2: extracting personal information from the collected image using the information extraction unit; extracting features from the collected image information using the feature extraction unit; and storing the extracted personal information and feature information using the data storage unit;

[0078] S3: using the information matching unit to match the personal information in the original database, and after the personal information is successfully matched, matching is performed with the corresponding feature information in the original database through the handheld object matching unit, the hand feature matching unit, and the angle matching unit;

[0079] S4: Real-time notification based on personal information and feature matching results in S3, and reminder for manual verification in case of mismatch;

[0080] S5: Intelligently update the database according to the verification results in S4.

[0081] In step S2, there are two ways to extract personal information from the captured image. One is to automatically identify and read the personal information in the QR code through an image input device and an image scanning device. The other is to pre-process the image and use the dlib library to identify the face of the image and find the eye feature points and calculate the inclination angle of the eye, rotate the photo and extract the head portrait and convert the image into a binary image, and then use the pytesseract library to identify the text information on the image, and extract the personal information through the print function; the feature extraction includes handheld object feature extraction, hand feature extraction, and angle extraction, and the handheld object feature extraction and hand feature extraction are extracted using a convolutional neural network; the binary image is a special grayscale image, that is, the grayscale value of any pixel is 0 or 255, representing black and white respectively; the dlib library, pytesseract library, print function, and convolutional neural network are all known technologies and are not described in detail here; the image pre-processing text image is separated and handed over to the recognition module for recognition, in order to eliminate useless information in the image, retain useful information, and simplify data to the maximum extent;

[0082] The angle extraction refers to identifying and extracting the handheld object in the captured image, and the camera of the image input device and the image scanning device is parallel to the extract to collect the image, and the collected picture of the extract is processed and stored as a standard picture in the standard object storage unit of the database, and the vertex of the upper left corner of the standard picture is used as the starting point to assign four vertices ABCD in a clockwise order, and a coordinate system is established with point D as the origin, DC as the positive direction of the x-axis, and DA as the positive direction of the y-axis, and the intersection of AC and BD is connected as the center point M, the coordinates of point M are represented by (a, b), and the coordinates of point A are represented by (x A ,y A ), the coordinates of point C are represented by (x C ,y C ) indicates that its center point The center point M is used as the standard reference point, and the lengths of DC and DA are measured to be L1 and L2 respectively;

[0083] The extract of the handheld object in the collected image is labeled with corresponding vertices according to the standard picture in the rectangular coordinate system. A corresponds to E, B corresponds to F, C corresponds to G, and D corresponds to H. A rectangular coordinate system is established with H as the origin, the horizontal direction as the x-axis, and the position perpendicular to the horizontal direction as the y-axis. The lengths l1 and l2 of HE and HG are measured, and l1 and l2 are compared with L1 and L2. If they are not equal, the extracted image is enlarged in the same proportion so that l1 = L1 and l2 = :2. EG and FH are connected and the intersection is taken as the center point Q. The coordinates of point Q are represented by (m,n), and the coordinates of point E are represented by (x E ,y E ), the coordinates of point G are represented by (xG ,y G ) indicates that its center point Put the M and Q coordinates into the same rectangular coordinate system with the origin at O, then ∠MOQ is the extracted tilt angle of the handheld object.

[0084] The feature matching in step S3 is based on region matching, and the angle matching is used to compare the angle information in the original database with the angle information extracted from the captured image. Put into the set K = {θ1,θ2,……,θ n-1}, and average the data in the K sets to get the optimal angle δ. The angle matching rate is: Said ε is the minimum maximum error of the angle matching rate;

[0085] when Indicates that the angle matching is successful;

[0086] When other situations occur, it means that the angle matching fails.

[0087] Step S4 notifies the matching result of S3 in real time. If a mismatch occurs, the staff is reminded to perform manual verification. After the manual verification is passed, step S5 updates the database, for example, by voice reminder, pop-up window prompt, or flashing warning light.

[0088] It can perform accurate and rapid identity recognition in large-scale scenarios, and effectively identify impersonators, effectively reducing safety hazards and avoiding dangerous accidents.

[0089] Example:

[0090] The data acquisition module collects images of the ID card held by the person to be tested through a mobile phone during large-scale health testing, and transmits the collected data to the data analysis and processing module;

[0091] The data analysis and processing module is to pre-process the collected image through the feature extraction unit, identify the face on the ID card using the dlib library, find the eye feature points and calculate the eye tilt angle, rotate the photo and extract the head portrait and convert the image into a binary image, and then use the pytesseract library to identify the text information on the picture, use the print function to extract the name, gender, address, date of birth, and ID number information on the ID card, and use the convolutional neural network to extract the handheld object feature and hand feature;

[0092] For the collected standard image, take the vertex in the upper left corner as the starting point and assign four vertices ABCD in clockwise order, and establish a coordinate system with point D as the origin, DC as the positive direction of the x-axis, and DA as the positive direction of the y-axis. Connect the intersection of AC and BD as the center point M. The coordinates of point M are represented by (a, b), the coordinates of point A are represented by (0, 2), and the coordinates of point C are represented by (4, 0). The center point The center point M is used as the standard reference point, and the lengths of DC and DA are measured to be L1=4 and L2=2 respectively;

[0093] The extract from the handheld object in the collected image is labeled with corresponding vertices according to the standard picture in the rectangular coordinate system. A corresponds to E, B corresponds to F, C corresponds to G, and D corresponds to H. A rectangular coordinate system is established with H as the origin, the horizontal direction as the x-axis, and the position perpendicular to the horizontal direction as the y-axis. The lengths of HE and HG are measured to be l1=1 and l2=2, and l1 and l2 are compared with L1 and L2. The extracted image is enlarged in proportion so that l1=L1=2 and l2=L2=4. EG and FH are connected and the intersection is taken as the center point Q. The coordinates of point Q are expressed as (m, n). The coordinates of point E are measured as Indicated by, the coordinates of point G are expressed as Indicates that its center point Put the M and Q coordinates into the same rectangular coordinate system with the origin at O, then ∠MOQ is the extracted tilt angle of the handheld object.

[0094] The set K in the original database is {-13.331°, -13.332°, -13.333°}, and the data in the K set are averaged to obtain the optimal angle. The angle matching rate is:

[0095] when Indicates that the angle matching is successful.

[0096] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0097] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or replace some of the technical features therein by equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A data distribution management system in a large-scale scenario, characterized by: The data distribution management system includes a data acquisition module, a data analysis and processing module, a feature comparison and judgment module, a verification notification module, and a database update module; The data acquisition module is used to transmit the data collected in large-scale scenarios to the data analysis and processing module; The data analysis and processing module is used to analyze and process the collected data; The feature comparison and judgment module is used to compare and match the feature information of the collected information with the corresponding feature information in the original database; The verification notification module is used to notify the staff of the comparison and matching results of the feature information in real time, and to perform on-site verification of mismatches; The database update module stores the collected data into the original database; The output end of the data acquisition module is connected to the input end of the data analysis and processing module; the output end of the data analysis and processing module is connected to the input end of the feature comparison and judgment module; the output end of the feature comparison module is connected to the input end of the verification notification module, and the output end of the verification notification module is connected to the input end of the database update module; The data acquisition module acquires data through an image input device and an image scanning device; The data analysis and processing module includes an information extraction unit, a feature extraction unit, and a data storage unit; The information extraction unit is used to extract useful personal information from the collected data, and the personal information includes name, gender, home address, ID number, headshot, phone number, and QR code information; The feature extraction unit is used to extract features from the collected data; The data storage unit is used to store the extracted personal information and characteristic information; The output end of the information extraction unit is connected to the input end of the feature extraction unit; the output end of the feature extraction unit is connected to the input end of the data storage unit; The feature comparison and judgment module includes an information matching unit, a handheld object matching unit, a hand feature matching unit, and an angle matching unit; The information matching unit matches the collected personal information with the personal information in the original database; The handheld object matching unit matches the handheld object information in the collected image with the corresponding handheld object information in the original database; The hand feature matching unit is used to match the collected hand features of the image with the corresponding feature information in the original database; The angle matching unit is used to match the angle of the handheld object with the angle of the corresponding handheld object in the original database; The output end of the personal information matching unit is connected to the input ends of the handheld object matching unit, the hand feature matching unit, and the angle matching unit.

2. The data distribution management system in a large-scale scenario according to claim 1, characterized in that: The verification notification module is used to notify the staff of the result of feature matching in real time and perform on-site verification of mismatches.

3. The data distribution management system in a large-scale scenario according to claim 2, characterized in that: The database update module is used to save the collected correct information to the corresponding position of the original database to complete the update of the database.

4. A data distribution management method in a large-scale scenario, the data distribution management method comprising the following steps: S1: Use image input equipment and image scanning equipment to collect image information and transmit it to the data analysis and processing module in real time; S2: extracting personal information from the captured image using an information extraction unit; Using a feature extraction unit to extract features from the collected image information; Using the data storage unit to store the extracted personal information and characteristic information; S3: using the information matching unit to match the personal information in the original database, and after the personal information is successfully matched, matching is performed with the corresponding feature information in the original database through the handheld object matching unit, the hand feature matching unit, and the angle matching unit; S4: Real-time notification based on personal information and feature matching results in S3, and reminder for manual verification in case of mismatch; S5: Intelligently update the database according to the verification result in S4; There are two ways to extract personal information from the captured image in step S2. One is to automatically identify and read the personal information in the QR code through an image input device and an image scanning device. The other is to pre-process the image and use the dlib library to identify the face of the image and find the eye feature points and calculate the inclination angle of the eye. The photo is rotated and the head portrait is extracted and the image is converted into a binary image. The pytesseract library is used to identify the text information on the image, and the personal information is extracted through the print function. The feature extraction includes handheld feature extraction, hand feature extraction, and angle extraction. The handheld feature extraction and hand feature extraction are extracted using a convolutional neural network. The angle extraction refers to identifying and extracting the handheld object in the captured image, and collecting images by arranging the camera of the image input device and the image scanning device in parallel with the extract, storing the collected image of the extract as the standard image in the standard object storage unit of the database after processing, and assigning four vertices ABCD in a clockwise order from the vertex in the upper left corner of the standard image as the starting point, and establishing a coordinate system with point D as the origin, DC as the positive direction of the x-axis, and DA as the positive direction of the y-axis, connecting the intersection of AC and BD as the center point M, and the coordinates of point M are represented by (a, b), and the coordinates of point A are represented by (x A ,y A ), the coordinates of point C are represented by (x C ,y C ) indicates that its center point , the center point M is taken as the standard reference point, and the lengths of DC and DA are measured to be L1 and L2 respectively; The extract of the handheld object in the collected image is numbered according to the corresponding position of the standard picture in the rectangular coordinate system. A corresponds to E, B corresponds to F, C corresponds to G, and D corresponds to H. A rectangular coordinate system is established with H as the origin, the horizontal direction as the x-axis, and the position perpendicular to the horizontal direction as the y-axis, and the lengths of HE and HG are measured. , and Compare with L1 and L2. If they are not equal, enlarge the extracted image by the same proportion. , connect EG and FH and take the intersection point as the center point Q. The coordinates of point Q are represented by (m,n), and the coordinates of point E are represented by (x E ,y E ) represents the coordinates of point G, and (x G ,y G ) indicates that its center point , put the coordinates of M and Q into the same rectangular coordinate system with the origin at O, then is the extracted handheld object tilt angle, .

5. The data distribution management method in a large-scale scenario according to claim 4 is characterized in that: The feature matching in step S3 is based on region matching, and the angle matching is used to compare the angle information in the original database with the angle information extracted from the captured image. Add to collection And calculate the average of the data in K sets to get the optimal angle , the angle matching rate is: ; when , indicating that the angle matching is successful; When other situations occur, it means that the angle matching fails.

6. The data distribution management method in a large-scale scenario according to claim 4 is characterized in that: Step S4 notifies the matching result of S3 in real time. If a mismatch occurs, the staff is reminded to perform manual verification. After the manual verification is passed, step S5 updates the database.

Images