Merchant abnormal license detection method and device, equipment and medium

Abstract

The invention relates to the technical field of image detection, and provides a commercial tenant abnormal license detection method and device, equipment and a medium, and the method comprises the steps: obtaining a plurality of to-be-detected pictures; detecting each to-be-detected picture by using a pre-trained open domain target detection model to determine the license type of the to-be-detected picture; when it is determined that the license type of the to-be-detected picture is a business license, filtering the to-be-detected picture by using a preset license elimination rule to determine whether the business license is an abnormal picture; and when it is determined that the business license is not the abnormal picture, performing OCR identification on the business license to obtain an identification result. Through the technical scheme of the invention, the accuracy and speed of business license detection are improved, the problem of information identification input errors caused by low picture quality is reduced, and the business efficiency is improved.

Description

[Technical field]

[0001] The present application relates to the field of image detection technology, and in particular to a method, device, equipment and medium for detecting abnormal merchant certificates. [Background technology]

[0002] A business license is a certificate issued by the industrial and commercial administration to industrial and commercial enterprises and individual business operators to allow them to engage in certain production and business activities. It is also the company's logo. The business license is equivalent to an ID card for the company. When enterprises participate in some platform projects, they need to upload business license pictures. Currently, there is no image quality detection for business license pictures at the project upload port. Instead, only the size of the picture is limited. The correctness, security, clarity, angle, contrast and other qualities of the received business license pictures cannot be controlled in a limited way, which easily leads to the inability to effectively identify the successfully uploaded business license pictures, which can only be solved through manual intervention.

[0003] like Figure 1 As shown in the figure, in the face of a large amount of image information, the traditional verification method is to manually check whether the merchant's license belongs to the specified category. For example, the merchant system requires the upload of a business license, but in fact the account manager may upload pictures such as the legal person certificate of a public institution or the registration certificate of a social group legal person, which are similar in function and appearance to the business license. They may also upload pictures such as account opening permits and bank cards, which are far from the function and appearance of the business license, or even some life photos of the merchant, which are completely irrelevant to the requirements. In addition, some low-quality pictures will be uploaded, such as pictures with shadows or blurs, and the business license fields cannot be seen clearly due to the long shooting distance. Both types of abnormal pictures need to be manually identified by the auditor, and the business logic is used to determine whether to return them to the account manager for modification.

[0004] This method can easily affect the progress and efficiency of business processes, waste storage space, generate junk images, and seriously affect or delay the progress of normal business. [Summary of the invention]

[0005] The embodiments of the present application provide a method, device, equipment and medium for detecting abnormal business licenses of merchants, aiming to solve the technical problem in the related art of manually detecting business licenses, which affects or delays the progress of normal business.

[0006] In a first aspect, an embodiment of the present application provides a method for detecting abnormal merchant certificates, including:

[0007] Get multiple pictures to be detected;

[0008] Using a pre-trained open domain object detection model to detect each image to be detected to determine the type of certificate of the image to be detected;

[0009] When it is determined that the license type of the image to be detected is a business license, the image to be detected is filtered using a preset license rejection rule to determine whether the business license is an abnormal image;

[0010] When it is determined that the business license is not an abnormal picture, OCR recognition is performed on the business license to obtain a recognition result.

[0011] In one embodiment, optionally, the method further comprises:

[0012] Obtain business license images of each merchant from the merchant system;

[0013] Get other types of certificate images;

[0014] The initial open domain object detection model is trained using the business license images of the various merchants and other types of license images to obtain the open domain object detection model.

[0015] In one embodiment, optionally, the method further comprises:

[0016] When it is determined that the certificate type of the image to be detected is other types of certificates, determining whether to perform certificate recognition according to preset business requirements;

[0017] When it is determined that the image to be detected does not belong to any existing certificate type, manually selecting a certificate image of a known type;

[0018] The known type of certificate image is input into the open domain object detection model for training to add a new type of recognized certificate to the open domain object detection model.

[0019] In one embodiment, optionally, the image to be detected is filtered using a preset license rejection rule to determine whether the business license is an abnormal image, including:

[0020] Calculate the proportion of the license image in the entire image to be detected;

[0021] When the ratio value is less than a preset threshold, it is determined that the business license is an abnormal picture.

[0022] In one embodiment, optionally, the image to be detected is filtered using a preset certificate rejection rule, including:

[0023] Use the preset image to construct the Laplace variance filter;

[0024] Using a Laplace variance filter to perform difference calculation on all the images to be detected, to obtain a range of difference scores;

[0025] Determining a preset difference threshold value according to the range of the difference score;

[0026] For the to-be-detected pictures whose difference is lower than the preset difference threshold, pictures are eliminated.

[0027] In one embodiment, optionally, the method further comprises:

[0028] The business license is subjected to OCR recognition, and when no recognition result is obtained, it is determined that the business license is an abnormal image.

[0029] In one embodiment, optionally, the method further comprises:

[0030] When it is determined that the business license is an abnormal picture, the abnormal picture is sent to a preset terminal to review and determine whether the abnormal picture needs to be modified again.

[0031] In a second aspect, an embodiment of the present application provides a device for detecting abnormal merchant certificates, including:

[0032] An acquisition module is used to acquire multiple images to be detected;

[0033] A first determination module is used to detect each image to be detected using a pre-trained open domain object detection model to determine the type of certificate of the image to be detected;

[0034] A filtering module, for filtering the image to be detected using a preset license rejection rule to determine whether the business license is an abnormal image when it is determined that the license type of the image to be detected is a business license;

[0035] The recognition module is used to perform OCR recognition on the business license to obtain a recognition result when it is determined that the business license is not an abnormal image.

[0036] In one embodiment, optionally, the device further comprises:

[0037] The first image acquisition module is used to acquire the business license image of each merchant from the merchant system;

[0038] The second image acquisition module is used to acquire other types of certificate images;

[0039] The first training module is used to train the initial open domain object detection model using the business license images of the merchants and other types of license images to obtain the open domain object detection model.

[0040] In one embodiment, optionally, the device further comprises:

[0041] A second determination module is used to determine whether to perform certificate recognition according to preset business requirements when it is determined that the certificate type of the image to be detected is other types of certificates;

[0042] A selection module, used for manually selecting a certificate image of a known type when it is determined that the image to be detected does not belong to any existing certificate type;

[0043] The second training module is used to input the known type of certificate images into the open domain object detection model for training, so as to add new identification certificate types to the open domain object detection model.

[0044] In one embodiment, optionally, the filtering module includes:

[0045] A first calculation unit is used to calculate the proportion of the certificate image in the entire image to be detected;

[0046] The first determining unit is used to determine that the business license is an abnormal image when the ratio value is less than a preset threshold.

[0047] In one embodiment, optionally, the filtering module includes:

[0048] A construction unit, used for constructing a Laplace variance filter using a preset image;

[0049] A second calculation unit is used to perform difference calculation on all the pictures to be detected using a Laplace variance filter to obtain a range of difference scores;

[0050] A second determining unit, configured to determine a preset difference threshold value according to a range of the difference score;

[0051] The elimination unit is used to eliminate the images to be detected whose difference is lower than the preset difference threshold.

[0052] In one embodiment, optionally, the identification module is further used to:

[0053] The business license is subjected to OCR recognition, and when no recognition result is obtained, it is determined that the business license is an abnormal image.

[0054] In one embodiment, optionally, the device further comprises:

[0055] The sending module is used to send the abnormal picture to a preset terminal to review and determine whether the abnormal picture needs to be modified when it is determined that the business license is an abnormal picture.

[0056] In a third aspect, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above-mentioned method for detecting abnormal merchant certificates when executing the computer program.

[0057] In a fourth aspect, a computer-readable storage medium is provided, which stores a computer program. When the computer program is executed by a processor, the steps of the above-mentioned merchant abnormal certificate detection method are implemented.

[0058] In the scheme implemented by the above merchant abnormal license detection method, device, equipment and medium, multiple pictures to be detected can be obtained; each picture to be detected is detected using a pre-trained open domain target detection model to determine the license type of the picture to be detected; when it is determined that the license type of the picture to be detected is a business license, the picture to be detected is filtered using a preset license elimination rule to determine whether the business license is an abnormal picture; when it is determined that the business license is not an abnormal picture, the business license is subjected to OCR recognition to obtain a recognition result. In the present invention, by detecting the pictures to be detected through an open domain target detection model, not only can the business license be accurately detected, but also other license types can be learned, thereby improving the accuracy and speed of business license detection, and abnormal pictures of business licenses can be further eliminated to reduce the problem of information recognition and entry errors caused by low picture quality, thereby improving business efficiency.

Brief Description of the Drawings

[0059] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0060] Figure 1 A schematic flow chart of a merchant certificate detection method in the related art is shown.

[0061] Figure 2 A schematic flow chart of a method for detecting abnormal merchant credentials according to an embodiment of the present application is shown.

[0062] Figure 3 A schematic flow chart of step S203 in a method for detecting abnormal merchant credentials according to an embodiment of the present application is shown.

[0063] Figure 4 A schematic flow chart of step S203 in a method for detecting abnormal merchant credentials according to another embodiment of the present application is shown.

[0064] Figure 5A schematic flow chart of a method for detecting abnormal merchant credentials according to another embodiment of the present application is shown.

[0065] Fig. 6A A schematic flow chart of a method for detecting abnormal merchant credentials according to yet another embodiment of the present application is shown.

[0066] Figure 6B A specific flow chart of a method for detecting abnormal merchant credentials according to an embodiment of the present application is shown.

[0067] Figure 7 A block diagram of a device for detecting abnormal merchant credentials according to another embodiment of the present application is shown.

[0068] Figure 8 A block diagram of a computer device according to an embodiment of the present application is shown. [Specific implementation method]

[0069] In order to better understand the technical solution of the present application, the embodiments of the present application are described in detail below with reference to the accompanying drawings.

[0070] It should be clear that the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in the field without creative work are within the scope of protection of the present application.

[0071] The terms used in the embodiments of the present application are only for the purpose of describing specific embodiments, and are not intended to limit the present application. The singular forms "a", "said" and "the" used in the embodiments of the present application and the appended claims are also intended to include plural forms, unless the context clearly indicates other meanings.

[0072] In conjunction with the accompanying drawings, some embodiments of the present application are described in detail below. In the absence of conflict, the following embodiments and features in the embodiments can be combined with each other.

[0073] See also Figure 2 , Figure 2 A schematic flow chart of a method for detecting abnormal merchant credentials according to an embodiment of the present application is shown.

[0074] like Figure 2 As shown, the process of the merchant abnormal certificate detection method according to an embodiment of the present application includes:

[0075] Step S201, obtaining multiple images to be detected;

[0076] Step S202, using a pre-trained open domain object detection model to detect each image to be detected, so as to determine the type of the certificate of the image to be detected;

[0077] In this step, it is necessary to obtain license images from the merchant system to train the basic open domain object detection model. Since the number of images in the merchant system is large and complex, manual selection is time-consuming and laborious. A simple trick can be used - use the existing business license OCR (optical character recognition, using electronic equipment to scan, recognize and print the text in the image) recognition interface to complete the recognition of all images, and write a script to determine whether the response message of each image contains the recognition results of the main fields, such as name, operator name, business location, business scope, registration number or unified social credit code. If it is included, the image is likely to be a business license, otherwise it is most likely not a business license. In this way, the scope can be quickly narrowed down, and a preset number of business licenses, such as 200 to 500 business licenses, can be selected from the remaining images to train the open domain object detection model.

[0078] Step S203: When it is determined that the license type of the image to be detected is a business license, the image to be detected is filtered using a preset license rejection rule to determine whether the business license is an abnormal image;

[0079] like Figure 3 As shown, in one embodiment, optionally, step S203 includes:

[0080] Step S301, calculating the proportion of the certificate image in the entire image to be detected;

[0081] Step S302: When the ratio value is less than a preset threshold, it is determined that the business license is an abnormal image.

[0082] In this embodiment, the proportion of the license image in the entire image to be detected can be calculated. For example, if the proportion of the license image is less than 60%, the business license is determined to be an abnormal image, and the abnormal image is eliminated.

[0083] like Figure 4 As shown, in one embodiment, optionally, step S203 includes:

[0084] Step S401, constructing a Laplace variance filter using a preset image;

[0085] The core of the Laplace variance filter is the Laplace operator, which is used to measure the second-order derivative of an image, emphasizing image areas that contain rapid intensity changes. The Laplace operator is often used for edge detection. There is an assumption here that if an image contains high variance, then there will be a large range of responses in the image, including edges and non-edges, which represents a normal image. But if the variance of the image is very low, then the range of responses is small, indicating that the edges in the image are small. It is well known that the more blurred the image is, the less edge information it contains.

[0086] Step S402, using a Laplace variance filter to perform difference calculation on all images to obtain a range of difference scores;

[0087] The core of the Laplace operator is a filter, and its mathematical formula is as follows:

[0088]

[0089] In formula (1), Δsrc represents the original image. This formula means that the second-order derivative of the original image in the X and Y directions is converted into the form of a convolution kernel:

[0090]

[0091] The Laplace variance filter is to perform a convolution operation on the original image through the above convolution kernel, and then calculate the square root. The final result is that each image will obtain a value, that is, the difference in the present invention.

[0092] Step S403, determining a preset difference threshold value according to the range of the difference score;

[0093] Step S404: for the to-be-detected pictures whose difference is lower than the preset difference threshold, pictures are eliminated.

[0094] Specifically, a threshold for filtering low-quality images can be obtained through manual comparison, and the threshold can be used to eliminate these low-quality images.

[0095] In the present invention, after all images are calculated by the Laplace variance filter, the obtained scores range from 0 to 5000, and a threshold of 300 is obtained after manual confirmation, that is, it can be considered that when the score is less than 300, the quality of the image is low.

[0096] Step S204: When it is determined that the business license is not an abnormal image, perform OCR recognition on the business license to obtain a recognition result.

[0097] In this embodiment, after filtering through the Laplace variance filter, there may still be a few low-quality images among the remaining images. At this time, the OCR interface can be used for identification. If the recognition result lacks key fields, such as registration number or unified social credit code, business premises, then the image can be considered to be still a low-quality image. After manual review by the auditor, it is determined whether the image quality is qualified. If it is determined to be unqualified, it can be returned to the customer manager for modification and re-upload.

[0098] In the above embodiment, all images are detected using an open domain target detection model. If the business license is not detected, other certificates of known categories are manually selected, and the data flows back to the open domain target detection model. The open domain target detection algorithm DE-ViT is used to continue to add new recognition types to the open domain target detection model; at the same time, other images are returned to the customer manager for modification and re-upload. If a business license is detected, abnormal images are further eliminated by calculating the license-to-image ratio and constructing a Laplace variance filter. Finally, the remaining images can be identified using the business license OCR recognition interface. If the recognition result lacks key fields, the abnormal images can be eliminated again, and the remaining recognition results can be entered into the system. In this way, not only can the business license be accurately detected, but other types of certificates can also be learned, thereby improving the accuracy and speed of business license detection, and abnormal images of business licenses can be further eliminated, reducing the problem of information recognition and entry errors caused by low image quality, and improving business efficiency.

[0099] like Figure 5 As shown, in one embodiment, optionally, the method further includes:

[0100] Step S501, obtaining the business license image of each merchant from the merchant system;

[0101] Step S502, obtaining other types of certificate images;

[0102] Step S503 , using the business license images of the merchants and other types of license images to train the initial open domain object detection model to obtain the open domain object detection model.

[0103] In this embodiment, the initial open domain object detection model is trained using business license images of the merchants and other types of certificate images. The open domain object detection model can detect and classify more and more types of certificates by continuously learning the features of other types of certificates, thereby providing a basis for auditors to judge whether to send the images to the OCR recognition interface of other relevant certificates, or to send the images back to the account manager for modification and re-upload.

[0104] like Fig. 6AAs shown, in one embodiment, optionally, the method further includes:

[0105] Step S601, when it is determined that the certificate type of the image to be detected is other types of certificates, determining whether to perform certificate recognition according to preset business requirements;

[0106] Step S602, when it is determined that the image to be detected does not belong to any existing certificate type, manually select a certificate image of a known type;

[0107] Step S603: input the known type of certificate image into the open domain object detection model for training, so as to add a new type of certificate to be recognized by the open domain object detection model.

[0108] In this embodiment, the open domain object detection model is used to detect all images. If the business license is not detected, other licenses of known categories are manually selected, and the data is returned to the open domain object detection model. The open domain object detection algorithm DE-ViT is used to continue to add new recognition types to the open domain object detection model.

[0109] The training set and test set of traditional target detection are both for the same target. For example, if there are only apples in the training set, then if other targets such as watermelons appear in the test set, the model will not be able to detect them. However, the test set of open domain target detection can be different from the training set. It can remember the features of new targets in the test set, and then train the model through manual labeling or other methods to let the model remember the new target, so that it can correctly infer the next time it detects the new target.

[0110] DE-ViT is a detector for open domain object detection. It can mark new targets in the test set, and then manually input a small number of new target samples and corresponding category names to allow the model to train and learn the characteristics of new targets, thereby achieving the goal of continuously improving the model's detection capabilities.

[0111] The architecture of the DE-ViT detector model is based on a new region-propagation mechanism (RPN). The propagated region mask is converted into a bounding box through a learnable spatial integration layer. In this way, there is no need to train a classifier specifically for a new target. Instead, you only need to build relevant prototypes through a script and feed them into the RPN network. Finally, a detector that can detect new targets can be trained with a small number of samples.

[0112] Below is the process of training the DE-ViT model.

[0113] 1) Label 450 business licenses, obtain the labeled JSON files, and convert them into coco format (an internationally recognized target detection labeling format);

[0114] 2) Build prototypes, train the RPN network, and obtain an open domain object detection model;

[0115] 3) After obtaining a certain number of samples of the new target, they are also labeled and converted to coco format, and prototypes are constructed;

[0116] 4) Select the few-shot training task with the training task type of fsod (few-shot object detection) to train the DE-ViT detector for the new target;

[0117] 5) Repeat steps 3) and 4) to allow the DE-ViT detector model to continuously learn to recognize new targets.

[0118] The above steps can realize continuous optimization and iteration of the open domain object detection model. The model can be used to infer the detection frame to calculate the size of the certificate in the original image to determine whether the shooting distance is too far, and can also be used to detect other types of certificates, reducing the judgment time of business personnel and improving work efficiency.

[0119] In one embodiment, optionally, the method further comprises:

[0120] The business license is subjected to OCR recognition, and when no recognition result is obtained, it is determined that the business license is an abnormal image.

[0121] In this embodiment, OCR recognition is performed on the business license. If no recognition result is obtained, it is determined that the business license is an abnormal image, and subsequent operations are performed.

[0122] In one embodiment, optionally, the method further comprises:

[0123] When it is determined that the business license is an abnormal picture, the abnormal picture is sent to a preset terminal to review and determine whether the abnormal picture needs to be modified again.

[0124] In this embodiment, if it is determined that the business license is an abnormal image, the abnormal image can be sent to a preset terminal, such as a terminal of an auditor, and the auditor will review it and determine whether the image needs to be revised.

[0125] The merchant abnormal certificate detection method of the present invention needs to be deployed on the GPU server. The overall process is as follows: Figure 6BAs shown. First, you need to obtain license images from the merchant system to train the basic business license detection model. Since the number of images in the merchant system is large and complex, manual selection is time-consuming and laborious. Here you can use a simple trick - use the existing business license OCR (optical character recognition, using electronic equipment to scan, recognize and print the text in the image) recognition interface, recognize all the images, and write a script to determine whether the response message of each image contains the recognition results of the main fields, such as name, operator name, business location, business scope, registration number or unified social credit code. If it is included, the image is likely to be a business license, otherwise it is most likely not a business license. In this way, you can quickly narrow the scope and select 200 to 500 business licenses from the remaining images to train the business license detection model.

[0126] All images are tested using the business license detection model. If no business license is detected, other licenses of known categories are manually selected, and the data is returned to the business license detection model. The open domain object detection algorithm DE-ViT is used to continue to add new recognition types to the basic model. At the same time, other images are returned to the account manager for modification and re-upload. If a business license is detected, abnormal images are further eliminated by calculating the license ratio and constructing a Laplace variance filter. Finally, the remaining images can be identified using the business license OCR recognition interface. If the recognition result lacks key fields, the abnormal images can be eliminated again and the remaining recognition results can be entered into the system.

[0127] The business license detection model can classify more and more types of certificates by continuously learning the features of other types of certificates, thereby providing auditors with a basis for judgment - whether to send the image to the OCR recognition interface of other relevant certificates, or to send the image back to the customer manager for modification and re-upload.

[0128] The present invention also carried out merchant license verification in a certain area, with a total of 91,361 pictures. After selection, a total of 450 business licenses were used to train the basic detection model (150 horizontal versions, 150 vertical versions, and 150 two-column versions). Subsequently, only 10 to 30 pictures of each type of license are needed for training other types of license pictures.

[0129] Figure 7 A block diagram of a device for detecting abnormal merchant credentials according to another embodiment of the present application is shown.

[0130] like Figure 7 As shown, in a second aspect, an embodiment of the present application provides a merchant abnormal certificate detection device 70, comprising:

[0131] An acquisition module 71 is used to acquire multiple images to be detected;

[0132] A first determination module 72 is used to detect each image to be detected using a pre-trained open domain object detection model to determine the type of the certificate of the image to be detected;

[0133] The filtering module 73 is used to filter the image to be detected by using a preset license rejection rule to determine whether the business license is an abnormal image when it is determined that the license type of the image to be detected is a business license;

[0134] The recognition module 74 is used to perform OCR recognition on the business license to obtain a recognition result when it is determined that the business license is not an abnormal image.

[0135] In one embodiment, optionally, the device further comprises:

[0136] The first image acquisition module is used to acquire the business license image of each merchant from the merchant system;

[0137] The second image acquisition module is used to acquire other types of certificate images;

[0138] The first training module is used to train the initial open domain object detection model using the business license images of the merchants and other types of license images to obtain the open domain object detection model.

[0139] In one embodiment, optionally, the device further comprises:

[0140] A second determination module is used to determine whether to perform certificate recognition according to preset business requirements when it is determined that the certificate type of the image to be detected is other types of certificates;

[0141] A selection module, used for manually selecting a certificate image of a known type when it is determined that the image to be detected does not belong to any existing certificate type;

[0142] The second training module is used to input the known type of certificate images into the open domain object detection model for training, so as to add new identification certificate types to the open domain object detection model.

[0143] In one embodiment, optionally, the filtering module includes:

[0144] A first calculation unit is used to calculate the proportion of the certificate image in the entire image to be detected;

[0145] The first determining unit is used to determine that the business license is an abnormal image when the ratio value is less than a preset threshold.

[0146] In one embodiment, optionally, the filtering module includes:

[0147] A construction unit, used for constructing a Laplace variance filter using a preset image;

[0148] A second calculation unit, used for calculating the difference of the image to be detected by using the Laplace variance filter;

[0149] The second determining unit is configured to determine that the business license is an abnormal image when the difference is not within a preset variance range.

[0150] In one embodiment, optionally, the identification module is further used to:

[0151] The business license is subjected to OCR recognition, and when no recognition result is obtained, it is determined that the business license is an abnormal image.

[0152] In one embodiment, optionally, the device further comprises:

[0153] The sending module is used to send the abnormal picture to a preset terminal to review and determine whether the abnormal picture needs to be modified when it is determined that the business license is an abnormal picture.

[0154] In a third aspect, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above-mentioned method for detecting abnormal merchant certificates when executing the computer program.

[0155] In a fourth aspect, a computer-readable storage medium is provided, which stores a computer program. When the computer program is executed by a processor, the steps of the above-mentioned merchant abnormal certificate detection method are implemented.

[0156] It should be noted that those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the above-described device and each module can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here.

[0157] The above-mentioned device can be implemented in the form of a computer program. Figure 8 Runs on the computer device shown.

[0158] Figure 8 A block diagram of a computer device according to an embodiment of the present application is shown.

[0159] See also Figure 8 The computer device includes a processor, a memory and a network interface connected through a system bus, wherein the memory may include a storage medium and an internal memory.

[0160] The storage medium can store an operating system and a computer program. The computer program includes program instructions, and when the program instructions are executed, the processor can execute any of the infectious disease spatial prediction methods for multi-source data provided in the embodiments of the present application.

[0161] The processor is used to provide computing and control capabilities and support the operation of the entire computer equipment.

[0162] The internal memory provides an environment for the operation of the computer program in the storage medium. When the computer program is executed by the processor, the processor can execute any method. The storage medium can be non-volatile or volatile.

[0163] The network interface is used for network communication, such as sending assigned tasks, etc. Those skilled in the art will understand that Figure 8 The structure shown in the figure is only a block diagram of a part of the structure 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 certain components, or have a different arrangement of components.

[0164] It should be understood that the processor may be a central processing unit (CPU), and the processor may also be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. Among them, the general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0165] In addition, an embodiment of the present application provides a computer-readable storage medium storing computer-executable instructions, wherein the computer-executable instructions are used to execute the steps of implementing the above-mentioned merchant abnormal certificate detection method.

[0166] It should be noted that the above functions or steps that can be implemented by the computer-readable storage medium or electronic device can refer to the relevant description in the aforementioned method embodiment. To avoid repetition, they will not be described one by one here.

[0167] The above, in conjunction with the accompanying drawings, describes in detail the technical solution of the present application. Through the technical solution of the present application, the relevant operations of grayscale release can be integrated into the release system. Developers only need to perform integrated settings in the release system to enable the release system to call the deployment system to perform corresponding deployment for the grayscale release, thereby reducing the complexity of the grayscale release deployment work and improving the efficiency and reliability of the grayscale release.

[0168] It should be understood that the term "and / or" used in this article is only a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship.

[0169] It should be understood that, although the terms first, second, etc. may be used to describe the setting unit in the embodiments of the present application, these setting units should not be limited to these terms. These terms are only used to distinguish the setting units from each other. For example, without departing from the scope of the embodiments of the present application, the first setting unit may also be referred to as the second setting unit, and similarly, the second setting unit may also be referred to as the first setting unit.

[0170] The word "if" as used herein may be interpreted as "at the time of" or "when" or "in response to determining" or "in response to detecting", depending on the context. Similarly, the phrases "if it is determined" or "if (stated condition or event) is detected" may be interpreted as "when it is determined" or "in response to determining" or "when detecting (stated condition or event)" or "in response to detecting (stated condition or event)", depending on the context.

[0171] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0172] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of hardware plus software functional units.

[0173] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and 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-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0174] The embodiments described above are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included in the protection scope of the present invention.

Claims

1. A method for detecting abnormal merchant certificates, characterized in that: include: Get multiple pictures to be detected; Using a pre-trained open domain object detection model to detect each image to be detected to determine the type of certificate of the image to be detected; When it is determined that the license type of the image to be detected is a business license, the image to be detected is filtered using a preset license rejection rule to determine whether the business license is an abnormal image; When it is determined that the business license is not an abnormal picture, OCR recognition is performed on the business license to obtain a recognition result.

2. The merchant abnormal certificate detection method according to claim 1, characterized in that: The method further comprises: Obtain business license images of each merchant from the merchant system; Get other types of certificate images; The initial open domain object detection model is trained using the business license images of the various merchants and other types of license images to obtain the open domain object detection model.

3. The merchant abnormal certificate detection method according to claim 1, characterized in that: The method further comprises: When it is determined that the certificate type of the image to be detected is other types of certificates, determining whether to perform certificate recognition according to preset business requirements; When it is determined that the image to be detected does not belong to any existing certificate type, manually selecting a certificate image of a known type; The known type of certificate image is input into the open domain object detection model for training to add a new type of recognized certificate to the open domain object detection model.

4. The method for detecting abnormal merchant credentials according to claim 1, characterized in that: The image to be detected is filtered using a preset license rejection rule to determine whether the business license is an abnormal image, including: Calculate the proportion of the license image in the entire image to be detected; When the ratio value is less than a preset threshold, it is determined that the business license is an abnormal picture.

5. The merchant abnormal certificate detection method according to claim 1, characterized in that: The image to be detected is filtered using the preset certificate rejection rules, including: Use the preset image to construct the Laplace variance filter; Using a Laplace variance filter to perform difference calculation on all the images to be detected, to obtain a range of difference scores; Determining a preset difference threshold value according to the range of the difference score; For the to-be-detected pictures whose difference is lower than the preset difference threshold, pictures are eliminated.

6. The merchant abnormal certificate detection method according to claim 1, characterized in that: The method further comprises: The business license is subjected to OCR recognition, and when no recognition result is obtained, it is determined that the business license is an abnormal image.

7. The method for detecting abnormal merchant certificates according to any one of claims 1 to 6, characterized in that: The method further comprises: When it is determined that the business license is an abnormal picture, the abnormal picture is sent to a preset terminal to review and determine whether the abnormal picture needs to be modified again.

8. A device for detecting abnormal merchant certificates, characterized in that: include: An acquisition module is used to acquire multiple images to be detected; A first determination module is used to detect each image to be detected using a pre-trained open domain object detection model to determine the type of certificate of the image to be detected; A filtering module, for filtering the image to be detected using a preset license rejection rule to determine whether the business license is an abnormal image when it is determined that the license type of the image to be detected is a business license; The recognition module is used to perform OCR recognition on the business license to obtain a recognition result when it is determined that the business license is not an abnormal image.

9. A computer device, characterized in that: include: at least one processor; and, a memory communicatively coupled to the at least one processor; The memory stores instructions executable by the at least one processor, wherein the instructions are configured to execute the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: Computer executable instructions are stored, and the computer executable instructions are used to execute the method according to any one of claims 1 to 7.

Images