Qualification picture auditing method and device, computer equipment and storage medium

Abstract

The invention relates to a qualification picture auditing method and device, computer equipment and a storage medium. The method comprises the steps that a to-be-audited qualification picture, collected by a user, of a to-be-audited enterprise is acquired; identifying to-be-audited information associated with the to-be-audited enterprise in the related picture, wherein the to-be-audited information comprises an object included in the to-be-audited qualification picture, an object type of the object and text information; the text information comprises an enterprise name and address information; obtaining standard qualification data of the to-be-audited enterprise based on a preset enterprise information table; performing hierarchical structured processing on the standard qualification data to obtain structured multi-layer standard qualification information; the multi-layer standard qualification information comprises enterprise names and address information; and calculating the similarity between the to-be-audited information and the structured multi-layer standard qualification information to determine an audit result of the to-be-audited qualification picture of the to-be-audited enterprise. The similarity is calculated based on a computer, and the determination of an audit result is realized; the manual auditing cost is reduced, and the auditing process efficiency is improved.

Description

Technical Field

[0001] This application relates to the technical field of data processing, and in particular, to a method, device, computer device, storage medium, and computer program product for auditing qualification pictures. Background Art

[0002] When a merchant or enterprise applies for corresponding qualifications to a target institution, in order to improve service quality and do a good job in risk prevention and control, the target institution will send corresponding staff to conduct on-site verification of information such as the qualification certificates and office environment of the merchant or enterprise. During the on-site verification process, the staff needs to take pictures for evidence collection to achieve the collection of pictures of the merchant or enterprise's office environment, and then the collected pictures are verified through remote manual review to determine whether the evidentiary pictures are qualified.

[0003] However, it is difficult to ensure the timeliness of manual photo review. Unqualified photos may not be notified to the staff who collected the pictures until several days later, and the staff needs to make a second on-site visit for re-picture collection, resulting in a cumbersome and long verification work process, and also causing the merchant or enterprise to take too long to apply for relevant qualifications, which is not conducive to the business expansion of the merchant or enterprise. Summary of the Invention

[0004] Based on this, in view of the above technical problems, it is necessary to provide a method, device, computer device, computer-readable storage medium, and computer program product for auditing qualification pictures that can improve the efficiency of picture review.

[0005] In a first aspect, this application provides a method for auditing qualification pictures, including:

[0006] Obtain the to-be-audited qualification pictures of the to-be-audited enterprise collected by the user;

[0007] Identify the to-be-audited information associated with the to-be-audited enterprise in the to-be-audited qualification pictures, where the to-be-audited information includes the objects included in the to-be-audited qualification pictures, the object types of the objects, and the text information in the objects; wherein, the text information at least includes the enterprise name and address information of the to-be-audited enterprise;

[0008] Based on a preset enterprise information table, obtain the standard qualification data of the to-be-audited enterprise;

[0009] Perform hierarchical structuring on the standard qualification data to obtain structured multi-layer standard qualification information; wherein, the multi-layer standard qualification information at least includes the enterprise name and address information of the to-be-audited enterprise;

[0010] Calculate the similarity between the information to be reviewed and the structured multi-level standard qualification information, and determine the review result of the qualification picture to be reviewed of the enterprise to be reviewed based on the similarity.

[0011] In one embodiment, the determining the review result of the qualification picture to be reviewed of the enterprise to be reviewed based on the similarity includes:

[0012] When the result value of the similarity is less than the first preset threshold, determine that the review result of the qualification picture to be reviewed of the enterprise to be reviewed fails the review;

[0013] Show the difference information to the user; the difference information includes the information to be reviewed with differences and the corresponding standard qualification information.

[0014] In one embodiment, the determining the review result of the qualification picture to be reviewed of the enterprise to be reviewed based on the similarity includes:

[0015] When the result value of the similarity is greater than or equal to the first preset threshold, determine that the review result of the qualification picture to be reviewed of the enterprise to be reviewed passes the review; at the same time, end the review task of the qualification picture to be reviewed and save the qualification picture to be reviewed.

[0016] In one embodiment, the hierarchical structuring process of the standard qualification data to obtain the structured multi-level standard qualification information includes:

[0017] Input the standard qualification data into a preset language model;

[0018] Obtain the structured multi-level standard qualification information based on the output of the preset language model; wherein, each level of standard qualification information includes an attribute and the text information corresponding to the attribute;

[0019] The multiple attributes corresponding to the multi-level standard qualification information include at least two of road name, road number, building name, floor number, door number, and enterprise name.

[0020] In one embodiment, after obtaining the structured multi-level standard qualification information, it further includes:

[0021] Correct the multi-level standard qualification information based on the standard qualification data.

[0022] In one embodiment, before identifying the information to be reviewed related to the enterprise to be reviewed in the qualification picture to be reviewed, it further includes:

[0023] Judge the orientation of the qualification picture to be reviewed;

[0024] When the matching degree between the orientation of the qualification picture to be audited and the preset orientation is less than the preset matching degree threshold, rotate the qualification picture to be audited to the preset orientation.

[0025] In one embodiment, after identifying the object associated with the enterprise to be audited in the qualification picture to be audited, it further includes:

[0026] Judge the integrity of the object;

[0027] When the integrity of the object is less than the second preset threshold, prompt the user to re-collect the corresponding qualification picture to be audited.

[0028] In one embodiment, before obtaining the qualification picture to be audited of the enterprise to be audited collected by the user, it further includes:

[0029] Show the user the collection process of the qualification picture to be audited, so as to prompt the user to collect the qualification picture to be audited according to the collection process.

[0030] In one embodiment, the obtaining of the qualification picture to be audited of the enterprise to be audited collected by the user includes:

[0031] Obtain at least one of a road sign including a road name and / or a house number, a building including a building name, a floor sign including a floor number, a house number plate including a house number, a company qualification license plate, an office scene, and a business license collected by the user and associated with the enterprise to be audited as the qualification picture to be audited.

[0032] In one embodiment, the identifying of the information to be audited associated with the enterprise to be audited in the qualification picture to be audited includes:

[0033] Identify at least one of the road sign, the building, the floor sign, the house number plate, the company qualification license plate, the office scene, and the business license associated with the enterprise to be audited in the qualification picture to be audited.

[0034] In one embodiment, the calculating of the similarity between the information to be audited and the structured multi-layer standard qualification information includes:

[0035] Calculate the sub-similarity between the text information of at least one of the road sign, the building, the floor sign, the house number plate, the company qualification license plate, the office scene, and the business license and the corresponding standard qualification information respectively;

[0036] Based on multiple sub-similarities, calculate the similarity between the information to be audited and the structured multi-layer standard qualification information.

[0037] In a second aspect, the present application also provides an auditing device for qualification pictures, and the device includes:

[0038] A data acquisition module, configured to acquire the to-be-audited qualification pictures of the enterprise to be audited collected by the user;

[0039] A data processing module, configured to identify the to-be-audited information associated with the enterprise to be audited in the to-be-audited qualification pictures, where the to-be-audited information includes the objects included in the to-be-audited qualification pictures, the object types of the objects, and the text information in the objects; wherein, the text information at least includes the enterprise name and address information of the enterprise to be audited;

[0040] The data acquisition module is further configured to acquire the standard qualification data of the enterprise to be audited based on a preset enterprise information table;

[0041] The data processing module is further configured to perform hierarchical structured processing on the standard qualification data to obtain structured multi-level standard qualification information; wherein, the multi-level standard qualification information at least includes the enterprise name and address information of the enterprise to be audited;

[0042] A data calculation and result output module, configured to calculate the similarity between the to-be-audited information and the structured multi-level standard qualification information, and determine the auditing result of the to-be-audited qualification pictures of the enterprise to be audited based on the similarity.

[0043] In a third aspect, the present application also provides a computer device, including a memory and a processor, where the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0044] Acquire the to-be-audited qualification pictures of the enterprise to be audited collected by the user;

[0045] Identify the to-be-audited information associated with the enterprise to be audited in the to-be-audited qualification pictures, where the to-be-audited information includes the objects included in the to-be-audited qualification pictures, the object types of the objects, and the text information in the objects; wherein, the text information at least includes the enterprise name and address information of the enterprise to be audited;

[0046] Acquire the standard qualification data of the enterprise to be audited based on a preset enterprise information table;

[0047] Perform hierarchical structured processing on the standard qualification data to obtain structured multi-level standard qualification information; wherein, the multi-level standard qualification information at least includes the enterprise name and address information of the enterprise to be audited;

[0048] Calculate the similarity between the information to be reviewed and the structured multi - level standard qualification information, and determine the review result of the qualification picture to be reviewed of the enterprise to be reviewed based on the similarity.

[0049] Fourthly, the present application also provides a computer - readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented:

[0050] Obtain the qualification picture to be reviewed of the enterprise to be reviewed collected by the user;

[0051] Identify the information to be reviewed associated with the enterprise to be reviewed in the qualification picture to be reviewed. The information to be reviewed includes the objects included in the qualification picture to be reviewed, the object types of the objects, and the text information in the objects; wherein, the text information at least includes the enterprise name and address information of the enterprise to be reviewed;

[0052] Based on a preset enterprise information table, obtain the standard qualification data of the enterprise to be reviewed;

[0053] Perform hierarchical structuring on the standard qualification data to obtain structured multi - level standard qualification information; wherein, the multi - level standard qualification information at least includes the enterprise name and address information of the enterprise to be reviewed;

[0054] Calculate the similarity between the information to be reviewed and the structured multi - level standard qualification information, and determine the review result of the qualification picture to be reviewed of the enterprise to be reviewed based on the similarity.

[0055] Fifthly, the present application also provides a computer program product, including a computer program. When the computer program is executed by a processor, the following steps are implemented:

[0056] Obtain the qualification picture to be reviewed of the enterprise to be reviewed collected by the user;

[0057] Identify the information to be reviewed associated with the enterprise to be reviewed in the qualification picture to be reviewed. The information to be reviewed includes the objects included in the qualification picture to be reviewed, the object types of the objects, and the text information in the objects; wherein, the text information at least includes the enterprise name and address information of the enterprise to be reviewed;

[0058] Based on a preset enterprise information table, obtain the standard qualification data of the enterprise to be reviewed;

[0059] Perform hierarchical structuring on the standard qualification data to obtain structured multi - level standard qualification information; wherein, the multi - level standard qualification information at least includes the enterprise name and address information of the enterprise to be reviewed;

[0060] Calculate the similarity between the information to be reviewed and the structured multi-layer standard qualification information, and determine the review result of the qualification picture to be reviewed of the enterprise to be reviewed based on the similarity.

[0061] The above-mentioned qualification picture review method, device, computer device, storage medium and computer program product identify the information to be reviewed, such as the objects, object types and text information in the objects associated with the enterprise to be reviewed in the picture, and then calculate the similarity between the information to be reviewed and the standard qualification data of the enterprise to be reviewed that has been processed by hierarchical structuring. Based on the similarity calculation result value, it is determined whether the qualification picture to be reviewed of the enterprise to be reviewed collected is qualified. It can be seen that the review process of the qualification picture to be reviewed of the enterprise to be reviewed in this application is implemented by a computer, and there is no need to manually identify and verify the information in the picture one by one. Based on the result value of the similarity between the information to be reviewed calculated by the computer and the structured multi-layer standard qualification information, the review of whether the corresponding picture is qualified can be realized. Compared with manual review, it can improve the review efficiency of picture information, reduce the manual review cost, reduce the cost of the picture collector making a second visit, and is also beneficial to improving the process efficiency of enterprise qualification review. Brief Description of the Drawings

[0062] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the following will briefly introduce the drawings required for use in the description of the embodiments or related technologies. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0063] Figure 1 It is a schematic flowchart of the qualification picture review method in one embodiment;

[0064] Figure 2 It is a schematic flowchart of the qualification picture review method in another embodiment;

[0065] Figure 3 It is a structural block diagram of the qualification picture review device in one embodiment;

[0066] Figure 4 It is an internal structure diagram of a computer device in one embodiment. Detailed Embodiments

[0067] In order to make the purpose, technical solutions and advantages of the present application clearer, the following further details the present application in combination with the drawings and embodiments. It should be understood that the specific embodiments described here are only used to explain the present application and are not used to limit the present application.

[0068] This embodiment takes the application of this method to a terminal as an example. Among them, the terminal can be, but is not limited to, various personal computers, laptop computers, smart phones, tablet computers, and Internet of Things devices. The Internet of Things devices can be, for example, smart TVs and in-vehicle devices. It can be understood that this method can also be applied to a server. The server can be an independent server or a server cluster composed of multiple servers. It can also be applied to a system including a terminal and a server and implemented through the interaction between the terminal and the server.

[0069] Figure 1 For the flowchart of the method for reviewing qualification pictures in an embodiment, please refer to Figure 1 In this embodiment, the method for reviewing qualification pictures may include the following steps 101 to 105, where:

[0070] Step 101, obtain the qualification pictures to be reviewed of the enterprise to be reviewed collected by the user.

[0071] Specifically, since the method for reviewing qualification pictures provided in this application can be applied to a terminal, specifically, for example, to an application software or a small program in the terminal. For the collection of the qualification pictures to be reviewed of the enterprise to be reviewed, the picture collection function integrated in the application software or the small program can be directly used to collect pictures of the on-site environment, qualification signs, etc. of the enterprise to be reviewed, so as to enable the application software or the small program to obtain the qualification pictures to be reviewed of the enterprise to be reviewed. In addition, if the user has previously collected and stored the relevant qualification pictures to be reviewed of the enterprise to be reviewed based on the terminal, the pictures can be uploaded to the application software or the small program to enable the application software or the small program to obtain the qualification pictures to be reviewed of the enterprise to be reviewed.

[0072] Among them, the enterprise to be reviewed is an enterprise or merchant that has submitted a qualification application to a relevant institution. The qualification pictures to be reviewed are pictures of the actual business premises corresponding to the enterprise or merchant. The pictures may contain any information related to the enterprise to facilitate the relevant institution to review the qualification certificate and office site of the enterprise. At the same time, the pictures related to the qualification certificate and office site of the enterprise can also be used to review the authenticity and reliability of the materials submitted by the enterprise itself when applying for qualifications.

[0073] It should be noted that this application does not specifically limit the way for the terminal to obtain the qualification pictures to be reviewed. Whether it is real-time collection or uploading of stored pictures is acceptable, as long as the terminal can obtain the relevant pictures.

[0074] Step 102: Identify the information to be reviewed in the qualification picture to be reviewed that is associated with the enterprise to be reviewed. The information to be reviewed includes the objects contained in the qualification picture to be reviewed, the object types of the objects, and the text information in the objects. Among them, the text information includes at least the enterprise name and address information of the enterprise to be reviewed.

[0075] Specifically, after the terminal obtains the qualification picture to be reviewed, Step 102 can be executed to review the relevant picture. The content of the review includes identifying the objects in the qualification picture to be reviewed that are associated with the enterprise to be reviewed, such as road signs, signs, office buildings, building numbers, etc. related to the enterprise to be reviewed. To avoid the influence of sundries in the collected pictures on the review results, it is also necessary to identify the object types of the objects included in the pictures. For example, it is necessary to identify the object with the object type of road sign from a picture including a road sign, a vehicle, and a tree. For example, it is necessary to identify the object with the object type of the building name from a picture including the sky, an airplane, the exterior of a building, and the building name, etc., so as to improve the review accuracy of the qualification picture to be reviewed and also help improve the review efficiency of the information associated with the enterprise to be reviewed. Further, it is also necessary to identify the text information of the text included in the identified objects associated with the enterprise to be reviewed to obtain the content displayed in the relevant objects. For example, it is to obtain the specific road name in the road sign, the specific name of the enterprise in the sign, the specific building name of the office building, etc.

[0076] That is to say, Step 102 for identifying the information to be reviewed in the qualification picture to be reviewed that is associated with the enterprise to be reviewed includes identifying the objects in the qualification picture to be reviewed, the object types of the objects, and the text information in the objects, so as to screen out the text information in the picture that is associated with the qualification enterprise to be reviewed. Specifically, the text information such as the specific office address and specific name related to the enterprise to be reviewed can be screened out. Further, the information obtained from the on-site pictures can also be used to compare with the information provided by the enterprise to the relevant institutions to confirm the authenticity and accuracy of the information provided by the enterprise to the relevant institutions.

[0077] Step 103: Based on the preset enterprise information table, obtain the standard qualification data of the enterprise to be reviewed.

[0078] Specifically, when an enterprise applies for qualifications, the information it provides to the relevant institutions will be stored in the enterprise information table. The terminal can obtain the corresponding standard qualification data submitted by the enterprise itself based on the preset enterprise information table storing the information of the enterprise to be reviewed. The standard qualification data can be used to compare with the text information in the pictures collected by the staff on the enterprise site to verify the accuracy of the pictures collected by the staff.

[0079] Step 104: Perform hierarchical structuring on the standard qualification data to obtain structured multi-level standard qualification information; among them, the multi-level standard qualification information includes at least the enterprise name and address information of the enterprise to be audited.

[0080] Specifically, after obtaining the standard qualification data of the enterprise to be audited, step 104 can be executed to perform hierarchical structuring on the standard qualification data to obtain structured multi-level standard qualification information, so as to avoid the increase in the difficulty of information recognition and information comparison caused by the aggregation of various information.

[0081] It should be added that, for example, the address information in the standard qualification data of the enterprise to be audited is "No. 3, 4th Floor, First (Street) Second (Road)", but the words "(Street)" and "(Road)" in the brackets are not included in the standard qualification data. If the information in the road sign extracted from the picture to be audited for qualification is "First Street Second Road", directly comparing the similarity between "First Second" and "First Street Second Road", it will surely be determined that the similarity between the two is very low. However, according to common sense, the specific content referred to by these two addresses is the same. It's just that when the enterprise provides the enterprise address information to the relevant agency, a more colloquial address writing method is adopted, omitting the words in the brackets. To avoid the audit result from being incorrect due to this situation, hierarchical structuring is performed on the standard qualification data, so that the obtained structured multi-level standard qualification information after processing is, for example, "Street: First; Road Name: Second; Road Number: Third; Floor: 4"; this clearly matches "First" with "Street" and "Second" with "Road Name".

[0082] By processing the standard qualification data into structured multi-level standard qualification information and then comparing it with the text information extracted from the picture, the problem of incorrect audit results caused by text omission can be avoided, which is conducive to improving the accuracy and efficiency of the audit results.

[0083] Step 105: Calculate the similarity between the information to be audited and the structured multi-level standard qualification information, and determine the audit result of the picture to be audited for qualification of the enterprise to be audited based on the similarity.

[0084] Specifically, after identifying the objects and text information associated with the enterprise to be audited in the qualification picture to be audited through step 102, and obtaining the structured multi-layer standard qualification information corresponding to the standard qualification data through step 104, step 105 is executed to calculate the similarity between the information to be audited and the structured multi-layer standard qualification information. Here, the calculation of the similarity is implemented through the background algorithm of the terminal, and there is no need for manual review and judgment of the information to be audited and the structured multi-layer standard qualification information; then, based on the calculation result value of the similarity, the accuracy and matching degree of the obtained qualification picture to be audited are judged to obtain the audit result of the qualification picture to be audited; it can be seen that auditing the qualification picture through the terminal can improve the audit efficiency of the picture information compared with manual auditing, reduce the manual audit cost, and thus can reduce the cost of the second visit of the picture collection personnel, and is also conducive to improving the process efficiency of enterprise qualification audit.

[0085] In the above method for auditing qualification pictures, in this application, the information to be audited, such as the objects, object types, and text information in the objects associated with the enterprise to be audited in the picture, is identified, and then the similarity between the information to be audited and the standard qualification data of the enterprise to be audited after hierarchical structuring is calculated. Based on the calculation result value of the similarity, it is determined whether the qualification picture to be audited of the enterprise to be audited is qualified; it can be seen that the audit process of the qualification picture to be audited of the enterprise to be audited in this application is implemented by a computer, and there is no need to manually identify and verify the information in the picture one by one. Based on the result value of the similarity between the information to be audited calculated by the computer and the structured multi-layer standard qualification information, the audit of whether the corresponding picture is qualified can be realized; compared with manual auditing, it can improve the audit efficiency of the picture information, reduce the manual audit cost, reduce the cost of the second visit of the picture collection personnel, and is also conducive to improving the process efficiency of enterprise qualification audit.

[0086] Figure 2 For the flowchart of the method for auditing qualification pictures in another embodiment, please refer to Figure 2 In an exemplary embodiment, the content of determining the audit result of the qualification picture to be audited of the enterprise to be audited based on the similarity in step 105 executed above may specifically include step 151 and step 152, where:

[0087] Step 151, when the result value of the similarity is less than the first preset threshold, determine that the audit result of the qualification picture to be audited of the enterprise to be audited is not passed.

[0088] Step 152, display the difference information to the user; the difference information includes the information to be audited with differences and the corresponding standard qualification information.

[0089] Specifically, for the content in step 105 of determining the review result of the to-be-reviewed qualification picture of the to-be-reviewed enterprise based on similarity, since the result values of the calculated similarity are different, the obtained review results will also be different. Here, a possible review result is specifically obtained by executing step 151 and step 152. Among them, when executing step 151, if the result value of the similarity calculated between the to-be-reviewed information and the structured multi-layer standard qualification information is less than the first preset threshold, it indicates that there is a large difference between the to-be-reviewed information and the structured multi-layer standard qualification information. That is, the to-be-reviewed information contained in the to-be-reviewed qualification picture obtained by the terminal is very different from the standard qualification data originally provided by the enterprise itself in the preset enterprise table. At this time, either the to-be-reviewed qualification picture uploaded by the picture collection personnel to the terminal is incorrect, or the enterprise has provided too much untrue information. Based on this, this review is naturally unqualified. Taking the standard qualification data originally provided by the enterprise itself in the preset enterprise table as the standard, naturally the review result of the to-be-reviewed qualification picture of the to-be-reviewed enterprise is not passed.

[0090] When it is learned through executing step 151 that the review result of the to-be-reviewed qualification picture of the to-be-reviewed enterprise in this review is not passed, step 152 can be further executed to display the difference information to the user, that is, the content of the difference between the to-be-reviewed information and the structured multi-layer standard qualification information found during the review process, so that the staff providing the to-be-reviewed qualification picture can learn the reason why this review is not passed and which picture(s) have differences resulting in the non-passing of the review result. That is, the difference information includes the to-be-reviewed information with differences and the corresponding standard qualification information, so that the staff collecting relevant pictures on-site of the to-be-reviewed enterprise can know which picture(s) need to be re-collected and updated.

[0091] It should be added that the present application does not specifically limit the size of the first preset threshold, and the size of the first preset threshold can be selected according to the review requirements. For example, the size of the first preset threshold can be selected as 88%, 95%, 98%, etc.

[0092] Please continue to refer to Figure 2 In an exemplary embodiment, the content of determining the review result of the to-be-reviewed qualification picture of the to-be-reviewed enterprise based on similarity in the above-mentioned executed step 105 can be specifically executed as step 153, where:

[0093] Step 153, when the result value of the similarity is greater than or equal to the first preset threshold, determine that the review result of the to-be-reviewed qualification picture of the to-be-reviewed enterprise is passed; at the same time, end the review task of the to-be-reviewed qualification picture and save the to-be-reviewed qualification picture.

[0094] Specifically, for the content in step 105 of determining the review result of the to-be-reviewed qualification picture of the to-be-reviewed enterprise based on similarity, since the result values of the calculated similarity are different, the obtained review results will also be different. Here, a possible review result is specifically obtained by executing step 153. Specifically, when the result value of the similarity between the to-be-reviewed information and the structured multi-layer standard qualification information is greater than or equal to the first preset threshold, it indicates that the to-be-reviewed information is the same as the structured multi-layer standard qualification information, or there are only very minor differences, and these minor differences are not sufficient to affect the matching between the to-be-reviewed information in the to-be-reviewed qualification picture and the standard qualification information. At this time, it can be determined that the review result of the to-be-reviewed qualification picture of the to-be-reviewed enterprise is passed. The terminal can end the review task of the to-be-reviewed qualification picture and store the obtained to-be-reviewed qualification picture.

[0095] Optionally, the to-be-reviewed qualification picture that has passed the review can be stored in a matching manner with the standard qualification data of the to-be-reviewed enterprise, so that when the staff of relevant institutions query the standard qualification data of a certain enterprise later, they can also query the qualification pictures that have passed the review of this enterprise, so that when the staff learn the text information of this enterprise, they can view the relevant picture information of this enterprise, which is conducive to the staff to more clearly understand the qualification data and office environment and other information of this enterprise.

[0096] Please continue to refer to Figure 2 , in an exemplary embodiment, for the content of the above-mentioned step 104 of performing hierarchical structuring on the standard qualification data to obtain structured multi-layer standard qualification information, it can be specifically executed as step 141 and step 142, where:

[0097] Step 141, input the standard qualification data into a preset language model;

[0098] Step 142, obtain structured multi-layer standard qualification information based on the output of the preset language model; where each layer of standard qualification information includes an attribute and the text information corresponding to the attribute; the multiple attributes corresponding to the multi-layer standard qualification information include at least two of road name, road number, building name, floor number, door number, and enterprise name.

[0099] Specifically, for the content of performing hierarchical structuring on the standard qualification data in step 104 to obtain structured multi-level standard qualification information, it can be specifically implemented by performing step 141 and step 142. The content executed in step 141 is to input the standard qualification data into a pre-trained language model. Since this language model has been pre-trained, it can process the standard qualification data input into it. Then, step 142 is executed, and structured multi-level standard qualification information processed by the pre-trained language model can be output based on it. Specifically, in each layer of standard qualification information output by the pre-trained language model, it includes an attribute and the text information corresponding to this attribute. The multiple attributes corresponding to the multi-level standard qualification information include at least two of road name, road number, building name, floor number, door number, and enterprise name. For example, the four layers of standard qualification information of "Street: First; Road Name: Second; Road Number: Third; Floor: 4".

[0100] Please continue to refer to Figure 2 , in an exemplary embodiment, after obtaining the structured multi-level standard qualification information in step 104 executed above, step 106 can be further executed to correct the multi-level standard qualification information based on the standard qualification data.

[0101] Specifically, in the qualification picture review method provided in this application, step 106 can be further included between step 104 and step 105. When step 104 is executed, since the standard qualification data provided by an enterprise to a relevant institution may not be provided according to a preset rule. For example, the standard address information of a certain enterprise should be "First Street, Second Road, Third Number, 4th Floor", but what this enterprise provides to the relevant institution is "First, Second, Third Number, 4th Floor", omitting "Street" and "Road" in the middle. Then, there may be differences or errors in the multi-level standard qualification information output after the address information is processed by the pre-trained language model. Therefore, after step 104 is executed, step 106 is further executed to correct the multi-level standard qualification information based on the standard qualification data, so as to eliminate the problem of incorrect output information caused by the enterprise not providing information according to the preset rule, which is beneficial to improving the accuracy of enterprise qualification review.

[0102] It should be added that step 106 can be manually operated by a staff member. For example, by judging the information provided by the enterprise based on known information to identify missing content or typos in the information, and adjusting the inaccurate information in the multi-level standard qualification information based on this judgment to ensure the accuracy of the multi-level standard qualification information. Only in this way can it be ensured that the structured multi-level standard qualification information is 100% accurate when calculating the similarity between the information to be reviewed and the structured multi-level standard qualification information, and it can be determined whether the obtained qualification picture to be reviewed is qualified according to the calculated value of the similarity.

[0103] However, the present application does not limit that step 106 can only be performed manually. The terminal can also analyze the standard qualification data provided by the enterprise. Specifically, it can match and correct the standard qualification data provided by the enterprise in combination with the standardized naming of information such as streets, roads, and buildings by the state. In this way, not only can the accuracy of multi-level standard qualification information be ensured, but also the review efficiency of qualification pictures can be further improved, and the labor cost can be greatly reduced.

[0104] Please continue to refer to Figure 2 , in an exemplary embodiment, before performing step 102 to identify the information to be reviewed associated with the enterprise to be reviewed in the qualification picture to be reviewed, steps 107 and 108 may be further performed, where:

[0105] Step 107, determine the orientation of the qualification picture to be reviewed;

[0106] Step 108, when the matching degree between the orientation of the qualification picture to be reviewed and the preset orientation is less than the preset matching degree threshold, rotate the qualification picture to be reviewed to the preset orientation.

[0107] Specifically, after obtaining the qualification picture to be reviewed of the enterprise to be reviewed collected by the user through step 101, step 107 can be first performed to determine the orientation of each qualification picture to be reviewed, so as to avoid the situation that some information in the picture cannot be recognized or is recognized incorrectly due to the incorrect orientation of the qualification picture to be reviewed; then step 108 is performed. When it is recognized that the matching degree between the orientation of the qualification picture to be reviewed and the preset orientation is less than the preset matching degree threshold, it means that the orientation of the qualification picture to be reviewed is incorrect. At this time, the qualification picture to be reviewed can be rotated to the preset orientation to improve the review efficiency and accuracy rate of the text information included in the qualification picture to be reviewed in the subsequent step 102.

[0108] It should be added that when the matching degree between the orientation of the qualification picture to be reviewed and the preset orientation is the same, it means that the orientation of the obtained qualification picture to be reviewed is already correct, and there is no need to adjust the orientation of the qualification picture to be reviewed. Step 102 can be directly performed to identify the information to be reviewed in the qualification picture to be reviewed.

[0109] Please continue to refer to Figure 2 , in an exemplary embodiment, after performing step 102 to identify the information to be reviewed associated with the enterprise to be reviewed in the qualification picture to be reviewed, steps 109 and 110 may be further performed, where:

[0110] Step 109, judge the integrity of the object;

[0111] Step 110: When the integrity of the object is less than the second preset threshold, prompt the user to re-collect the corresponding qualification picture to be reviewed.

[0112] Specifically, after identifying the information to be reviewed associated with the enterprise to be reviewed in the qualification picture to be reviewed through Step 102, or after identifying the object associated with the enterprise to be reviewed in the information to be reviewed associated with the enterprise to be reviewed in the qualification picture to be reviewed through Step 102, Step 109 can be executed first to judge the integrity of the object in the identified qualification picture to be reviewed, so as to avoid the situation that objects such as road signs and shop signs in the qualification picture to be reviewed are not completely photographed, which is conducive to avoiding the situation that the text information identified from the object is incomplete; and then Step 110 is executed. When it is judged that the integrity of the object is less than the second preset threshold, that is, when it is judged that there are objects in the qualification picture to be reviewed that are not completely photographed, the user can be prompted through the terminal to re-collect the corresponding qualification picture to be reviewed, so that the complete object to be reviewed can be presented in the qualification picture to be reviewed, thereby improving the integrity of the text information identified from the object, which is conducive to improving the accuracy of the review result of the qualification picture.

[0113] It should be added that when it is judged that all the objects associated with the enterprise to be reviewed in the qualification picture to be reviewed are complete, there is no need to display a prompt message to the user, and the next step can be directly executed; of course, it is also possible to prompt the user that the qualification picture to be reviewed provided by them is complete and they can choose to continue with the subsequent review steps. This application does not make specific limitations on this.

[0114] Please continue to refer to Figure 2 , in an exemplary embodiment, before executing Step 101 to obtain the qualification picture to be reviewed of the enterprise to be reviewed collected by the user, Step 201 can be executed first to display the collection process of the qualification picture to be reviewed to the user, so as to prompt the user to collect the qualification picture to be reviewed in accordance with the collection process.

[0115] Specifically, the method for auditing qualification pictures provided by this application is specifically applied to an application or a mini-program in a terminal. Therefore, before the application or mini-program in the terminal executes step 101 to obtain the qualification picture to be audited of the enterprise to be audited collected by the user, step 201 can be executed first to display the collection process of the qualification picture to be audited to the user through the application or mini-program. For example, the qualification pictures to be provided are displayed to the user in sections. For instance, on the page for providing the qualification picture to be collected, from top to bottom, the content in the first area is "Building exterior view", which is used to prompt that after the user enters this menu, the qualification picture to be provided is the exterior photo of the office building of the enterprise to be audited. The content in the second area is "Enterprise doorplate", which is used to prompt that after the user enters this menu, the qualification picture to be provided is the photo of the enterprise license plate of the enterprise to be audited, so as to achieve the effect of prompting the user to collect the qualification picture to be audited according to the collection process; let the user collect or provide the corresponding qualification pictures to be audited through different menu entries. Collecting different qualification pictures to be audited in this way is also conducive to the subsequent extraction of text information from different types of objects, making the text information and its corresponding object type clearer, and facilitating the calculation accuracy of the similarity between the information to be audited and the structured multi-layer standard qualification information in the subsequent process.

[0116] Please continue to refer to Figure 2 , in an exemplary embodiment, the above-mentioned step 101 of obtaining the qualification picture to be audited of the enterprise to be audited collected by the user can be specifically selected to be executed as:

[0117] Obtain a photo of at least one of a road sign including a road name and / or a house number, a building including a building name, a floor sign including a floor number, a house number, a company qualification license plate, an office scene, and a business license, which is associated with the enterprise to be audited and collected by the user, as the qualification picture to be audited.

[0118] Specifically, the step 101 of obtaining the qualification picture to be audited of the enterprise to be audited collected by the user can be specifically to obtain a photo of at least one of the information such as a road sign including a road name and / or a house number, a building including a building name, a floor sign including a floor number, a house number, a company qualification license plate, an office scene, and a business license, which is associated with the enterprise to be audited and collected by the user, and use it as the qualification picture to be audited; the required categories of the qualification picture to be audited can be defined by the developers of relevant application software or mini-programs, and the user can be guided by executing step 201 before obtaining these pictures, so that the user can know which pictures need to be collected as the qualification picture to be audited.

[0119] It should be added that, in addition to the above-mentioned categories, the categories of the qualification pictures to be reviewed may further include other categories, and this application does not make specific limitations on this. The categories of the qualification pictures to be reviewed can be adjusted according to the review requirements.

[0120] Please continue to refer to Figure 2 , in an exemplary embodiment, step 102 performed above for identifying the information to be reviewed associated with the enterprise to be reviewed in the qualification picture to be reviewed can specifically be performed as follows:

[0121] Identify at least one of the road signs, buildings, floor numbers, house numbers, company qualification licenses, office scenes, and business licenses associated with the enterprise to be reviewed in the qualification picture to be reviewed.

[0122] Specifically, for step 102 to identify the information to be reviewed associated with the enterprise to be reviewed in the qualification picture to be reviewed, it can specifically be to identify at least one of the information such as road signs, buildings, floor numbers, house numbers, company qualification licenses, office scenes, and business licenses associated with the enterprise to be reviewed in the qualification picture to be reviewed; the information categories identified in step 102 can be determined according to the selection of the developers of relevant application software or mini-programs. For example, when step 201 is executed and it shows which categories of qualification pictures to be reviewed need to be collected by the user, then step 102 will review the information of those categories.

[0123] It should be added that, in addition to the above-mentioned categories, the categories of the qualification pictures to be reviewed may further include other categories, and this application does not make specific limitations on this. The categories of the qualification pictures to be reviewed can be adjusted according to the review requirements.

[0124] Please continue to refer to Figure 2 , in an exemplary embodiment, for the content of calculating the similarity between the information to be reviewed and the structured multi-level standard qualification information in step 105 performed above, it can specifically be performed as follows:

[0125] Calculate the sub-similarities between the text information of at least one of the road signs, buildings, floor numbers, house numbers, company qualification licenses, office scenes, and business licenses and the corresponding sub-standard qualification information respectively;

[0126] Based on multiple sub-similarities, calculate the similarity between the information to be reviewed and the structured multi-level standard qualification information.

[0127] Specifically, the calculation of the similarity between the information to be reviewed and the structured multi-level standard qualification information performed in step 105 can be specifically as follows: calculate the sub-similarity between the text information of at least one of road signs, buildings, floor numbers, door numbers, company qualification licenses, office scenes, and business licenses and the corresponding standard qualification information respectively; when step 201 is executed and it shows which categories of qualification pictures to be reviewed need to be collected by the user, then step 105 will calculate the similarity between the information of those categories and the corresponding standard qualification information respectively; for example, if step 201 guides the user to collect road signs, buildings, and floor numbers associated with an enterprise to be reviewed, then when step 105 is executed, the sub-similarity between the text information extracted from the road sign and the sub-level standard qualification information corresponding to the road sign can be calculated first, then the sub-similarity between the text information extracted from the building and the sub-level standard qualification information corresponding to the building can be calculated, and then the sub-similarity between the text information extracted from the floor number and the sub-level standard qualification information corresponding to the floor number can be calculated. Finally, based on these three sub-similarities, the similarity between the information to be reviewed and the structured multi-level standard qualification information can be calculated; this is beneficial to improving the accuracy of the review results of qualification pictures.

[0128] An implementation manner of the qualification picture review method provided by this application is that for the enterprise to be inspected, the collector takes pictures of the enterprise site according to the collection process guidance of the application or mini-program and submits the pictures to the system background for review; after the pictures are processed by the object detection module and the OCR (Optical Character Recognition) recognition module of the AI (Artificial Intelligence) vision model, the objects and the corresponding text information are returned as the next input. Among them, the enterprise site photos collected by the terminal include, but are not limited to, scene pictures such as road signs, building names, signs, door numbers, office scenes, and business licenses; further, for the information recognized from the scene pictures, it includes objects related to the enterprise to be reviewed, the object types corresponding to the objects, the text information in the objects, and in addition, it can further include the coordinate information of the objects in the pictures.

[0129] Among them, for the pictures of enterprise doorplates or road signs collected by users, after reading the pictures, the orientation of the pictures can be judged first. If the orientation is different from the preset orientation, the pictures can be rotated. Then, the quadrilateral detection can be performed on the objects in the pictures (such as road signs or doorplates). If the objects in the pictures are not quadrilaterals, image cropping and correction can be performed on them to adjust them into objects presented as quadrilaterals. Then, the text in the objects is detected. When it is detected that the objects contain text, the text in the objects is recognized. Then, the recognized text can be combined into text. After the text is combined, information extraction is performed on the combined text, such as extracting address segment information. Finally, the extracted text information is output, such as the text information corresponding to the text in the doorplate or road sign.

[0130] It should be added that due to perspective reasons, doorplates or road signs generally appear as quadrilaterals in images. Therefore, a quadrilateral detection algorithm is proposed, which can be used to accurately detect doorplates or road signs in images. Since the text in doorplates or road signs is generally parallel to the edges of doorplates or road signs, after correcting the doorplates or road signs, the text is horizontally arranged in the images, which is very beneficial to subsequent text detection and recognition. Due to problems such as different font sizes, large inter-character spaces, different font styles, and text chessboard layouts in the text of doorplates or road signs, the usual text line detection-based methods have poor effects. Therefore, the character-level detection algorithm CRAFT (Character Region Awareness For Text detection) is adopted, which can effectively avoid text misdetection and omission caused by unclear line breaks. After single-character detection, based on the rules that the main text is generally located in the middle of the doorplate or road sign and has the largest font size, the secondary text is generally located above, below, in front of, or behind the main text, and the text of the same level has the same font size and is neatly arranged, the single characters are combined into text lines and the primary and secondary are distinguished. Then, a text recognition algorithm based on CRNN (Convolutional Recurrent Neural Network) and CTC (Connectionist Temporal Classification) is used to recognize the main text line and the secondary text line, rather than using a single-character recognition model. The advantage is that line-based recognition can connect the semantic information before and after to improve the recognition effect, and at the same time, it can also avoid the problem of inaccurate single-character recognition caused by single-character adhesion.

[0131] After obtaining the object types and text information of the objects associated with the enterprise to be audited in the pictures associated with the enterprise to be audited, the target enterprise information provided by the enterprise itself is extracted, including the company name and the enterprise address. Then, the target enterprise information is input into the language model, and the language model parses the enterprise address and extracts keywords. Then, the hierarchical information of the enterprise address corresponding to the target enterprise information is output.

[0132] Finally, calculate the similarity between the text information in the picture and the hierarchical information of the enterprise address, and determine whether the qualification picture passes the review based on the similarity result value.

[0133] It should be added that by learning the structural pattern of the address through a language model, such as a certain building on a certain road in a certain district of a certain city, input the address string to match the keywords and the order of address elements of the address structure pattern, split out the hierarchical address elements, parse to obtain key information such as road name, road number, building name, floor number, house number, enterprise name, etc. After the address information is hierarchically structured, it is matched with the picture information to achieve higher review accuracy and recall rate. Among them, when there is information redundancy or missing in the enterprise address text string input to the language model, after the language model outputs the hierarchical information of the enterprise address, the hierarchical information can be supplemented and corrected with key text information.

[0134] It also should be added that calculate the text similarity between the "object type and text information" returned after artificial intelligence calculation and the "hierarchical information of enterprise address" returned after address standardization processing. When the calculated result similarity reaches a certain threshold, it is determined that the collected picture is consistent with the target enterprise information; if not, combined with the detected object information and text matching situation, output specific difference points to the user. Among the information output to the user, it can also include whether the object in the picture is complete and what the address matching rate is.

[0135] It also should be added that after obtaining the similarity calculation result, the system background of the relevant application software or applet will synchronize the verification result to the window end of the application software or applet. If the result is qualified, the collector can submit an order; if the picture information is inconsistent with the target enterprise information, the collector will be prompted to confirm and correct. In a normal network state, the entire verification duration can be controlled within 30 seconds, and the collector can respond directly on-site, avoiding a second visit.

[0136] In addition, the user can also make remarks on the on-site situation photographed in the relevant application software or applet, and this remark behavior can be executed together with the photo collection process; for example, when the staff cannot find the floor sign at the enterprise site, then there is no way to provide the corresponding photo, or there is no text information about the floor where the enterprise is located in the provided photo, and relevant remarks can be submitted to the application software or applet.

[0137] In addition, the method for auditing qualification pictures provided in this application can be extended to other scenarios for auditing based on target objects and key texts. Through algorithm iteration and rule optimization, it can meet the different requirements of enterprise inspections. The current solution supports automatic auditing of enterprise POI (Point of Information) information, automatic auditing of business licenses, automatic auditing of interior office photos, automatic auditing of legal person group photos, and other projects.

[0138] It should be understood that although the steps in the flowcharts involved in the above-described embodiments are shown in sequence according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless there is a clear description in this article, the execution of these steps has no strict order limit, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-described embodiments may include multiple steps or multiple stages. These steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be executed alternately or in turn with at least a part of other steps or steps or stages in other steps.

[0139] Based on the same inventive concept, the embodiments of this application also provide an auditing device for qualification pictures for implementing the above-mentioned auditing method for qualification pictures. The solution provided by this device to solve the problems of cumbersome workflow, long process, and low auditing efficiency in manually verifying the pictures taken on-site by staff is similar to the solution described in the above method. Therefore, the specific limitations in one or more of the following embodiments of the auditing device for qualification pictures can refer to the limitations on the auditing method for qualification pictures in the above text, and will not be elaborated here.

[0140] Figure 3 For the structural block diagram of the auditing device for qualification pictures in an embodiment, please refer to Figure 1 、 Figure 2 Refer to Figure 3 In an exemplary embodiment, as shown in Figure 3 Figure 3 , an auditing device 200 for qualification pictures is provided. The device includes: a data acquisition module 81, a data processing module 82, and a data calculation and result output module 83, where:

[0141] The data acquisition module 81 is used to acquire the to-be-audited qualification pictures of the enterprise to be audited collected by the user;

[0142] A data processing module 82, which is used to identify the information to be audited associated with the enterprise to be audited in the qualification picture to be audited. The information to be audited includes the objects contained in the qualification picture to be audited, the object types of the objects, and the text information in the objects. Among them, the text information includes at least the enterprise name and address information of the enterprise to be audited.

[0143] The data acquisition module 81 is further used to obtain the standard qualification data of the enterprise to be audited based on a preset enterprise information table.

[0144] The data processing module 82 is further used to perform hierarchical structuring on the standard qualification data to obtain structured multi-layer standard qualification information. Among them, the multi-layer standard qualification information includes at least the enterprise name and address information of the enterprise to be audited.

[0145] A data calculation and result output module 83, which is used to calculate the similarity between the information to be audited and the structured multi-layer standard qualification information, and determine the audit result of the qualification picture to be audited of the enterprise to be audited based on the similarity.

[0146] Specifically, the data acquisition module 81 is used to collect the qualification picture to be audited of the enterprise to be audited. It can directly collect pictures of information such as the on-site environment and qualification signs of the enterprise to be audited based on the picture collection function integrated in the application software or applet, so as to realize the acquisition of the qualification picture to be audited of the enterprise to be audited by the application software or applet. In addition, if the user has previously collected and stored the relevant qualification pictures to be audited of the enterprise to be audited based on the terminal, the pictures can be uploaded to the application software or applet to realize the acquisition of the qualification picture to be audited of the enterprise to be audited by the application software or applet.

[0147] Among them, the enterprise to be audited is an enterprise or merchant that has submitted a qualification application to the relevant institution. The qualification picture to be audited is a picture of the actual business premises corresponding to the enterprise or merchant. The picture may contain any information related to the enterprise, so as to facilitate the relevant institution to audit the qualification certificate and office site of the enterprise, and to audit the authenticity and reliability of the materials submitted by the enterprise itself when applying for the qualification.

[0148] It should be noted that this application does not specifically limit the way for the terminal to obtain the qualification picture to be audited. Whether it is real-time collection or uploading of stored pictures, as long as the terminal can obtain the relevant pictures.

[0149] The data processing module 82 is used to review relevant pictures after the terminal obtains the pictures of qualifications to be reviewed. The content of the review includes identifying the objects associated with the enterprise to be reviewed in the pictures of qualifications to be reviewed, such as road signs, signs, office buildings, building numbers, etc. related to the enterprise to be reviewed. In order to avoid the influence of sundries in the collected pictures on the review results, it is also necessary to identify the object types of the objects included in the pictures. For example, it is necessary to identify the object with the object type of road sign from the pictures including road signs, vehicles, and trees. For example, it is necessary to identify the object with the object type of building name from the pictures including the sky, airplanes, the exterior of the building, and the building name, etc., so as to improve the review accuracy of the pictures of qualifications to be reviewed and also help improve the review efficiency of the information associated with the enterprise to be reviewed. Further, it is also necessary to identify the text information of the text included in the identified objects associated with the enterprise to be reviewed, so as to obtain the content displayed in the relevant objects. For example, it is to obtain the specific road name in the road sign, the specific name of the enterprise in the sign, the specific name of the office building, etc.

[0150] That is to say, the data processing module 82 is used to identify the information to be reviewed associated with the enterprise to be reviewed in the pictures of qualifications to be reviewed, including the identification of the objects, object types, and text information in the pictures of qualifications to be reviewed, so as to screen out the text information associated with the enterprise with qualifications to be reviewed in the pictures. Specifically, the text information such as the specific office address and specific name related to the enterprise to be reviewed can be screened out. These information can be used to compare with the information provided by the enterprise to relevant institutions to confirm the authenticity and accuracy of the information provided by the enterprise to relevant institutions.

[0151] When an enterprise applies for qualifications, the information about itself provided to relevant institutions will be stored in the enterprise information table. The data acquisition module 81 is also used to obtain the standard qualification data submitted by the corresponding enterprise itself from the preset enterprise information table storing the information of the enterprise to be reviewed. The standard qualification data can be used to compare with the text information in the pictures collected by the staff on the enterprise site to verify the accuracy of the pictures collected by the staff.

[0152] After obtaining the standard qualification data of the enterprise to be reviewed, the data processing module 82 is also used to perform hierarchical structured processing on the standard qualification data to obtain structured multi-layer standard qualification information, so as to avoid the increase in the difficulty of information identification caused by the aggregation of various information.

[0153] It should be added that, for example, the address information in the standard qualification data of the enterprise to be audited is "No. 3, 4th Floor, First (Street) Second (Road)", but the words "Street" and "Road" in the brackets are not included in the standard qualification data. If the information on the road sign extracted from the picture of the qualification to be audited is "First Street Second Road", directly comparing the similarity between "First Second" and "First Street Second Road", it will surely be determined that the similarity between the two is very low. However, according to common sense, the specific content referred to by these two addresses is the same. It's just that when the enterprise provides the enterprise address information to the relevant agency, it uses a more colloquial way of writing the address and omits the words in the brackets. To avoid errors in the audit results caused by this situation, the standard qualification data is processed in a hierarchical and structured manner, so that the structured multi-level standard qualification information obtained after processing is, for example, "Street: First; Road Name: Second; Road Number: Third; Floor: 4"; this clearly matches "First" with "Street" and "Second" with "Road Name".

[0154] By processing the standard qualification data into structured multi-level standard qualification information and then comparing it with the text information extracted from the picture, the problem of incorrect audit results caused by text omission can be avoided, which is beneficial to improving the accuracy and efficiency of the audit results.

[0155] The data calculation and result output module 83 is used to calculate the similarity between the information to be audited and the structured multi-level standard qualification information. Here, the calculation of similarity is implemented through the background algorithm of the terminal, and there is no need for manual review and judgment of the information to be audited and the structured multi-level standard qualification information; then, based on the calculated similarity result value, the accuracy and matching degree of the picture of the qualification to be audited obtained are judged to obtain the audit result of the picture of the qualification to be audited; it can be seen that by auditing the qualification picture through the terminal, compared with manual auditing, it can improve the audit efficiency of the picture information, reduce the manual audit cost, and thus can reduce the cost of the second visit of the picture collection personnel, and is also beneficial to improving the process efficiency of enterprise qualification audit.

[0156] In an exemplary embodiment, the data calculation and result output module 83 is used to determine the audit result of the picture of the qualification to be audited for the enterprise to be audited based on the similarity. Specifically, the data calculation and result output module 83 is used to determine that the audit result of the picture of the qualification to be audited for the enterprise to be audited is not passed when the result value of the similarity is less than the first preset threshold; then, the difference information is displayed to the user; the difference information includes the information to be audited with differences and the corresponding standard qualification information. Specifically, reference can be made to Figure 2 and the above description about Figure 2 .

[0157] In an exemplary embodiment, the data calculation and result output module 83 is used to determine the review result of the to-be-reviewed qualification picture of the to-be-reviewed enterprise based on the similarity. Specifically, when the result value of the similarity is greater than or equal to the first preset threshold, the data calculation and result output module 83 determines that the review result of the to-be-reviewed qualification picture of the to-be-reviewed enterprise is passed; meanwhile, the review task of the to-be-reviewed qualification picture is ended, and the to-be-reviewed qualification picture is saved. For details, reference can be made to Figure 2 , and the above description about Figure 2 .

[0158] In an exemplary embodiment, the data processing module 82 is used to perform hierarchical structuring on the standard qualification data to obtain structured multi-layer standard qualification information. Specifically, the data processing module 82 is used to input the standard qualification data into a preset language model; then, based on the output of the preset language model, structured multi-layer standard qualification information is obtained; wherein, each layer of standard qualification information includes an attribute and the corresponding text information; the multiple attributes corresponding to the multi-layer standard qualification information include at least two of road name, road number, building name, floor number, door number, and enterprise name. For details, reference can be made to Figure 2 , and the above description about Figure 2 .

[0159] In an exemplary embodiment, after obtaining the structured multi-layer standard qualification information, the data processing module 82 is further used to correct the multi-layer standard qualification information based on the standard qualification data. For details, reference can be made to Figure 2 , and the above description about Figure 2 .

[0160] In an exemplary embodiment, before the data processing module 82 is used to identify the to-be-reviewed information associated with the to-be-reviewed enterprise in the to-be-reviewed qualification picture, it is also used to judge the orientation of the to-be-reviewed qualification picture; when the matching degree between the orientation of the to-be-reviewed qualification picture and the preset orientation is less than the preset matching degree threshold, the to-be-reviewed qualification picture is rotated to the preset orientation. For details, reference can be made to Figure 2 , and the above description about Figure 2 .

[0161] In an exemplary embodiment, before the data processing module 82 is used to identify the to-be-reviewed information associated with the to-be-reviewed enterprise in the to-be-reviewed qualification picture, it is also used to judge the integrity of the object; when the integrity of the object is less than the second preset threshold, the user is prompted to re-collect the corresponding to-be-reviewed qualification picture. For details, reference can be made to Figure 2 , and the above description about Figure 2 .

[0162] In an exemplary embodiment, the review device 200 for qualification pictures further includes an interaction module, which is used to display the acquisition process of the qualification pictures to be reviewed to the user, so as to prompt the user to acquire the qualification pictures to be reviewed according to the acquisition process. For details, refer to Figure 2 , and the above description about Figure 2 .

[0163] In an exemplary embodiment, the data acquisition module 81 is used to acquire the qualification pictures to be reviewed of the enterprise to be reviewed collected by the user. Specifically, the data acquisition module 81 is used to acquire at least one of the pictures of road signs including road names and / or house numbers, buildings including building names, floor signs including floor numbers, house numbers, company qualification licenses, office scenes, and business licenses associated with the enterprise to be reviewed collected by the user, as the qualification pictures to be reviewed. For details, refer to Figure 2 , and the above description about Figure 2 .

[0164] In an exemplary embodiment, the data processing module 82 is used to identify the information to be reviewed associated with the enterprise to be reviewed in the qualification pictures to be reviewed. Specifically, the data processing module 82 is used to identify at least one of the road signs, buildings, floor signs, house numbers, company qualification licenses, office scenes, and business licenses associated with the enterprise to be reviewed in the qualification pictures to be reviewed. For details, refer to Figure 2 , and the above description about Figure 2 .

[0165] In an exemplary embodiment, the data calculation and result output module 83 is used to calculate the similarity between the information to be reviewed and the structured multi-level standard qualification information. Specifically, the data calculation and result output module 83 is used to calculate the sub-similarities between the text information of at least one of the road signs, buildings, floor signs, house numbers, company qualification licenses, office scenes, and business licenses and the corresponding standard qualification information respectively; based on multiple sub-similarities, calculate the similarity between the information to be reviewed and the structured multi-level standard qualification information. For details, refer to Figure 2 , and the above description about Figure 2 .

[0166] Each module in the above-mentioned review device 200 for qualification pictures can be implemented in whole or in part by software, hardware, and their combination. The above-mentioned modules can be embedded in the processor of the computer device in the form of hardware or independent of it, or stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above-mentioned modules.

[0167] Figure 4The internal structure diagram of a computer device in an embodiment. In an exemplary embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as shown in Figure 4 the figure. The computer device includes a processor, a memory, an input / output interface, a communication interface, a display unit, and an input device. Among them, the processor, the memory, and the input / output interface are connected through a system bus. The communication interface, the display unit, and the input device are connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and external devices. The communication interface of the computer device is used to communicate with external terminals in a wired or wireless manner. The wireless manner can be implemented through WIFI, a mobile cellular network, NFC (Near Field Communication), or other technologies. The computer program, when executed by the processor, implements a method for reviewing qualification pictures. The display unit of the computer device is used to form a visually visible picture, which can be a display screen, a projection device, or a virtual reality imaging device. The display screen can be a liquid crystal display screen or an electronic ink display screen. The input device of the computer device can be a touch layer covering the display screen, or a button, a trackball, or a touchpad provided on the outer shell of the computer device, or an external keyboard, touchpad, or mouse, etc.

[0168] Those skilled in the art can understand that Figure 4 the structure shown in the figure is only a block diagram of some structures related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine some components, or have different component arrangements.

[0169] Based on the same inventive concept, the present application also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it implements the aforementioned method for reviewing qualification pictures. The method for reviewing qualification pictures is any one of the methods for reviewing qualification pictures mentioned in the embodiments of the present application. For related embodiments, reference may be made to the above text.

[0170] Based on the same inventive concept, the present application also provides a computer program product, including a computer program. When the computer program is executed by a processor, it implements the aforementioned method for reviewing qualification pictures. The method for reviewing qualification pictures is any one of the methods for reviewing qualification pictures mentioned in the embodiments of the present application. For related embodiments, reference may be made to the above text.

[0171] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in this application are all information and data that have been authorized by the user or fully authorized by all parties, and the collection, use, and processing of relevant data need to comply with relevant regulations.

[0172] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, database, or other medium used in the embodiments provided in this application can include at least one of non-volatile and volatile memories. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc. The databases involved in the embodiments provided in this application can include at least one of relational databases and non-relational databases. Non-relational databases can include distributed databases based on blockchain, etc., and are not limited thereto. The processors involved in the embodiments provided in this application can be general-purpose processors, central processors, graphics processors, digital signal processors, programmable logic devices, data processing logics based on quantum computing, etc., and are not limited thereto.

[0173] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered to be within the scope described in this specification.

[0174] The above-described embodiments merely represent several implementation manners of the present application. The description thereof is relatively specific and detailed, but it should not be construed as a limitation to the patent scope of the present application. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present application, several modifications and improvements can still be made, and these all fall within the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the appended claims.

Claims

1. A method for reviewing qualification pictures, characterized in that, Including: Obtain the to-be-reviewed qualification pictures of the enterprise to be reviewed collected by the user; Identify the to-be-reviewed information associated with the enterprise to be reviewed in the to-be-reviewed qualification pictures, where the to-be-reviewed information includes the objects contained in the to-be-reviewed qualification pictures, the object types of the objects, and the text information in the objects; among them, the text information includes at least the enterprise name and address information of the enterprise to be reviewed; Based on a preset enterprise information table, obtain the standard qualification data of the enterprise to be reviewed; Perform hierarchical structuring on the standard qualification data to obtain structured multi-level standard qualification information; among them, the multi-level standard qualification information includes at least the enterprise name and address information of the enterprise to be reviewed; Calculate the similarity between the to-be-reviewed information and the structured multi-level standard qualification information, and determine the review result of the to-be-reviewed qualification pictures of the enterprise to be reviewed based on the similarity.

2. The method for reviewing qualification pictures according to claim 1, characterized in that, The determining the review result of the to-be-reviewed qualification pictures of the enterprise to be reviewed based on the similarity includes: In the case where the result value of the similarity is less than a first preset threshold, determine that the review result of the to-be-reviewed qualification pictures of the enterprise to be reviewed is not passed; Display the difference information to the user; the difference information includes the to-be-reviewed information with differences and the corresponding standard qualification information.

3. The method for reviewing qualification pictures according to claim 1, characterized in that, The determining the review result of the to-be-reviewed qualification pictures of the enterprise to be reviewed based on the similarity includes: In the case where the result value of the similarity is greater than or equal to the first preset threshold, determine that the review result of the to-be-reviewed qualification pictures of the enterprise to be reviewed is passed; at the same time, end the review task of the to-be-reviewed qualification pictures and save the to-be-reviewed qualification pictures.

4. The method for reviewing qualification pictures according to claim 1, characterized in that, The performing hierarchical structuring on the standard qualification data to obtain structured multi-level standard qualification information includes: Input the standard qualification data into a preset language model; Obtain structured multi-level standard qualification information based on the output of the preset language model; among them, each level of standard qualification information includes an attribute and the text information corresponding to the attribute; The multiple attributes corresponding to the multi-level standard qualification information include at least two of road name, road number, building name, floor number, door number, and enterprise name.

5. The method for reviewing qualification pictures according to claim 1 or 4, characterized in that, After obtaining the structured multi-level standard qualification information, it further includes: Correct the multi-level standard qualification information based on the standard qualification data.

6. The method for reviewing qualification pictures according to any one of claims 1-4, characterized in that, Before identifying the to-be-reviewed information associated with the enterprise to be reviewed in the to-be-reviewed qualification pictures, it further includes: Judge the orientation of the to-be-reviewed qualification pictures; In the case where the matching degree between the orientation of the to-be-reviewed qualification pictures and the preset orientation is less than the preset matching degree threshold, rotate the to-be-reviewed qualification pictures to the preset orientation.

7. The method for reviewing qualification pictures according to any one of claims 1-4, characterized in that, After identifying the object associated with the enterprise to be reviewed in the to-be-reviewed qualification pictures, it further includes: Judge the integrity of the object; In the case where the integrity of the object is less than a second preset threshold, prompt the user to re-collect the corresponding to-be-reviewed qualification pictures.

8. The method for reviewing qualification pictures according to any one of claims 1-4, characterized in that, Before obtaining the to-be-reviewed qualification picture of the enterprise to be reviewed collected by the user, it further includes: Showing the collection process of the to-be-reviewed qualification picture to the user to prompt the user to collect the to-be-reviewed qualification picture according to the collection process.

9. The method for reviewing qualification pictures according to any one of claims 1-4, characterized in that, The obtaining of the to-be-reviewed qualification picture of the enterprise to be reviewed collected by the user includes: Obtaining at least one of a road sign including a road name and / or a house number, a building including a building name, a floor sign including a floor number, a house number, a company qualification license plate, an office scene, and a business license associated with the enterprise to be reviewed collected by the user as the to-be-reviewed qualification picture.

10. The method for reviewing qualification pictures according to claim 9, characterized in that, The identifying of the to-be-reviewed information associated with the enterprise to be reviewed in the to-be-reviewed qualification picture includes: Identifying at least one of the road sign, the building, the floor sign, the house number, the company qualification license plate, the office scene, and the business license associated with the enterprise to be reviewed in the to-be-reviewed qualification picture.

11. The method for reviewing qualification pictures according to claim 10, characterized in that, The calculating of the similarity between the to-be-reviewed information and the structured multi-layer standard qualification information includes: Calculating the sub-similarity between the text information of at least one of the road sign, the building, the floor sign, the house number, the company qualification license plate, the office scene, and the business license and the corresponding standard qualification information respectively; Calculating the similarity between the to-be-reviewed information and the structured multi-layer standard qualification information based on the multiple sub-similarities.

12. A device for reviewing qualification pictures, characterized in that, The device includes: A data acquisition module for obtaining the to-be-reviewed qualification picture of the enterprise to be reviewed collected by the user; A data processing module for identifying the to-be-reviewed information associated with the enterprise to be reviewed in the to-be-reviewed qualification picture, where the to-be-reviewed information includes the object included in the to-be-reviewed qualification picture, the object type of the object, and the text information in the object; wherein, the text information at least includes the enterprise name and address information of the enterprise to be reviewed; The data acquisition module is further configured to obtain the standard qualification data of the enterprise to be reviewed based on a preset enterprise information table; The data processing module is further configured to perform hierarchical structuring processing on the standard qualification data to obtain structured multi-layer standard qualification information; wherein, the multi-layer standard qualification information at least includes the enterprise name and address information of the enterprise to be reviewed; A data calculation and result output module for calculating the similarity between the to-be-reviewed information and the structured multi-layer standard qualification information and determining the review result of the to-be-reviewed qualification picture of the enterprise to be reviewed based on the similarity.

13. A computer device, including a memory and a processor, the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 11.

14. A computer-readable storage medium, on which a computer program is stored, characterized in that, When the computer program is executed by the processor, it implements the steps of the method according to any one of claims 1 to 11.

15. A computer program product, including a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the method according to any one of claims 1 to 11.

Images