CN111898602A

CN111898602A - Certificate number area identification method, device and equipment in image

Info

Publication number: CN111898602A
Application number: CN202010794013.6A
Authority: CN
Inventors: 王宇涛; 鲁自恒; 秦铭
Original assignee: Agree Technology Co ltd
Current assignee: Agree Technology Co ltd
Priority date: 2020-08-10
Filing date: 2020-08-10
Publication date: 2020-11-06
Anticipated expiration: 2040-08-10
Also published as: CN111898602B

Abstract

The invention discloses a certificate number area identification method in an image, which can firstly determine a user certificate area corresponding to a user certificate in the image to be identified after the image to be identified is obtained, and then determine a certificate number area corresponding to the certificate number of the user certificate based on the user certificate area. Therefore, the user certificate area can be identified from the image to be identified, and the area characteristic of the certificate number can be identified in the user certificate area according to the area characteristic of the certificate number, so that the user does not need to accurately and manually mark the certificate number area of the user certificate, and the user can determine the certificate number area corresponding to the certificate number only by placing the user certificate in the acquisition area corresponding to the image to be identified, so that the input complexity of the image area corresponding to the certificate number is reduced, the identification accuracy of the certificate number area is improved, and the efficiency of determining the certificate number area in the image to be identified is improved.

Description

Certificate number area identification method, device and equipment in image

Technical Field

The invention relates to the technical field of computers, in particular to a method, a device and equipment for identifying a certificate number area in an image.

Background

Text appearing in natural scenes is an important source of information. Such as billboards, traffic signs, words on various certificates, etc., which contain explicit semantic information to provide necessary instructions and reminders to people. Understanding and analyzing the scene content can be achieved if detection and recognition of the text can be achieved. With the advancement of human science and technology, the use of machines to detect and understand text in scenes has become a necessary trend. With the development of image processing technology and artificial intelligence technology, people have higher and higher requirements on the intellectualization and accuracy of text recognition.

At present, the traditional identification mode of the certificate number area adopted by the user certificate processing counter is generally as follows: in the process that a client uses a machine for self-service handling, the client is required to operate the user certificate through special equipment according to the photoacoustic prompt so as to read the certificate number into the system. For example, the user is required to manually mark the image area corresponding to the voucher number in the user voucher image, and the device can read the voucher number. However, in the conventional user credential identification method, the user needs to manually and accurately determine the image area corresponding to the credential number, and if the user has a presentation deviation in the process of labeling the credential number area in the user credential image or manually and incorrectly labels the credential number area, the device cannot acquire the credential number area or the acquired credential number area is incomplete, and thus cannot determine the credential number, and at this time, the user needs to repeatedly label the credential number area of the user credential. Thus, the traditional mode is easy to cause the problem that the input complexity of the voucher number area is increased due to unfamiliarity or improper operation of the user, and further the input efficiency of the voucher number area is low.

Therefore, how to improve the input efficiency and the input accuracy of the image area corresponding to the voucher number is a technical difficulty to be solved at present.

Disclosure of Invention

The invention provides a certificate number area identification method and device in an image, and aims to realize the identification.

In a first aspect, the present invention provides a method for identifying a voucher number area in an image, the method comprising:

a method for identifying a voucher number area in an image, the method comprising:

acquiring an image to be identified, wherein the image to be identified comprises an image of a user certificate;

determining a user certificate area corresponding to the user certificate in the image to be identified;

and determining a certificate number area corresponding to the certificate number of the user certificate based on the user certificate area.

In a second aspect, the present invention provides an apparatus for identifying a voucher number area in an image, the apparatus comprising:

the device comprises an image acquisition unit, a recognition unit and a recognition unit, wherein the image acquisition unit is used for acquiring an image to be recognized, and the image to be recognized comprises an image of user credentials;

the first determining unit is used for determining a user certificate area corresponding to the user certificate in the image to be identified;

and the second determining unit is used for determining a certificate number area corresponding to the certificate number of the user certificate based on the user certificate area.

In a third aspect, the invention provides a readable medium comprising executable instructions, which when executed by a processor of an electronic device, cause the electronic device to perform the method according to any one of the first aspect.

In a fourth aspect, the present invention provides an electronic device, comprising a processor and a memory storing execution instructions, wherein when the processor executes the execution instructions stored in the memory, the processor performs the method according to any one of the first aspect.

According to the technical scheme, after the image to be identified is obtained, the image to be identified comprises the image of the user certificate; a user credential area corresponding to the user credential in the image to be recognized may be determined first, and then a credential number area corresponding to a credential number of the user credential may be determined based on the user credential area. Therefore, the user certificate area can be identified from the image to be identified, and the area characteristic of the certificate number can be identified in the user certificate area according to the area characteristic of the certificate number, so that the scheme provided by the application does not need a user to accurately and manually mark the certificate number area of the user certificate, and the certificate number area corresponding to the certificate number can be determined only by placing the user certificate in the acquisition area corresponding to the image to be identified, so that the input complexity of the image area corresponding to the certificate number is reduced, the identification accuracy of the certificate number area is improved, and the efficiency of determining the certificate number area in the image to be identified is improved.

Further effects of the above-mentioned unconventional preferred modes will be described below in conjunction with specific embodiments.

Drawings

In order to more clearly illustrate the embodiments or prior art solutions of the present invention, the drawings used in the description of the embodiments or prior art will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments described in the present invention, and it is obvious for a person skilled in the art to obtain other drawings based on these drawings without paying creative efforts.

FIG. 1 is a block diagram of an exemplary application scenario provided in an embodiment of the present invention;

fig. 2 is a flowchart illustrating a method for identifying a voucher number area in an image according to an embodiment of the present invention;

fig. 3a is a schematic diagram of an image to be recognized according to an embodiment of the present invention;

FIG. 3b is a diagram illustrating a user credential area according to an embodiment of the present invention;

FIG. 3c is a schematic diagram of a gray scale image corresponding to a user credential area according to an embodiment of the present invention;

fig. 3d is a schematic diagram of an edge detection image corresponding to a user credential area according to an embodiment of the present invention;

fig. 3e is an illustration diagram of a binary image corresponding to a user credential area (a blurred binary image of a bank card credential) according to an embodiment of the present invention;

FIG. 3f is a schematic diagram of an edge brightness statistic image corresponding to a user credential area according to an embodiment of the present invention;

FIG. 3g is a diagram illustrating a voucher number area according to an embodiment of the present invention;

FIG. 4 is a diagram illustrating the main direction of a user credential area according to an embodiment of the present invention;

fig. 5 is a schematic structural diagram of a device for identifying a voucher number area in an image according to an embodiment of the present invention;

fig. 6 is a schematic structural diagram of an electronic device according to an embodiment of the present invention.

Detailed Description

In order to make the objects, technical solutions and advantages of the present invention more apparent, the technical solutions of the present invention will be described in detail and completely with reference to the following embodiments and accompanying drawings. It is to be understood that the described embodiments are merely exemplary of the invention, and not restrictive of the full scope of the invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

In the prior art, the traditional identification mode of the certificate number area adopted by the user certificate processing counter is generally as follows: in the process that a client uses a machine for self-service handling, the client is required to operate the user certificate through special equipment according to the photoacoustic prompt so as to read the certificate number into the system. For example, if the user credential is a bank card and the credential number is a bank card number, the device may read the credential number only by manually marking an image area corresponding to the bank card number in the collected image including the bank card. However, in the conventional user credential identification method, the user needs to manually and accurately determine the image area corresponding to the credential number, and if the user has a representation deviation in the process of marking the credential number area in the user credential image or manually and incorrectly marks the credential number area, the device cannot acquire the credential number area or the acquired credential number area is incomplete, and thus cannot determine the credential number, and at this time, the user needs to repeatedly mark the credential number area of the user credential. Thus, the traditional mode easily causes the problem that the input complexity of the voucher number area is increased due to unfamiliar or improper operation of the user, and further causes the input efficiency of the voucher number area to be low. Therefore, how to improve the input efficiency and the input accuracy of the image area corresponding to the voucher number is a technical difficulty to be solved urgently at present. .

In order to solve the problem, the invention provides a certificate number area identification method in an image, which is used for acquiring an image to be identified, wherein the image to be identified comprises an image of a user certificate; a user credential area corresponding to the user credential in the image to be recognized may be determined first, and then a credential number area corresponding to a credential number of the user credential may be determined based on the user credential area. Therefore, the user certificate area can be identified from the image to be identified, and the area characteristic of the certificate number can be identified in the user certificate area according to the area characteristic of the certificate number, so that the scheme provided by the application does not need a user to accurately and manually mark the certificate number area of the user certificate, and can determine the certificate number area corresponding to the certificate number only by placing the user certificate in the acquisition area corresponding to the image to be identified, thereby not only reducing the input complexity of the image area corresponding to the certificate number, but also improving the identification accuracy of the certificate number area, and further improving the efficiency of determining the certificate number area in the image to be identified.

For example, embodiments of the present invention may be applied to the scenario shown in FIG. 1. In the scene, a camera 101 and a terminal device 102 are included; the terminal device 102 may be a terminal having a data processing function, for example, a terminal device such as a smart phone, a tablet computer, a desktop computer, and a notebook computer, and it can be understood that the terminal device 102 may also be other terminal devices such as a server, which is not limited in this embodiment. It should be noted that, in this embodiment, the camera 101 may be disposed on the terminal device 102, or may be disposed as an independent image capturing device.

In this scenario, the camera 101 may first acquire an image to be recognized, where the image to be recognized includes an image of a user credential; after the terminal device 102 receives the to-be-identified head portrait sent by the camera 101, the terminal device 101 may first determine a user credential area corresponding to the user credential in the to-be-identified image, and then may determine, based on the user credential area, a credential number area corresponding to a credential number of the user credential. Therefore, the user certificate area can be identified from the image to be identified, and the area characteristic of the certificate number can be identified in the user certificate area according to the area characteristic of the certificate number, so that the scheme provided by the application does not need a user to accurately and manually mark the certificate number area of the user certificate, and the certificate number area corresponding to the certificate number can be determined only by placing the user certificate in the acquisition area corresponding to the image to be identified, so that the input complexity of the image area corresponding to the certificate number is reduced, the identification accuracy of the certificate number area is improved, and the efficiency of determining the certificate number area in the image to be identified is improved.

It is to be understood that, in the above application scenario, although some actions of the embodiment of the present application are described as being performed by the camera 101 and some actions are performed by the terminal device 102, in other embodiments, all actions may be performed by the terminal device 102. The present application is not limited to the execution of the main body, and may be configured to execute the operations disclosed in the embodiments of the present application.

It should be noted that the above application scenarios are only shown for the convenience of understanding the present application, and the embodiments of the present application are not limited in any way in this respect. Rather, embodiments of the present application may be applied to any scenario where applicable.

Various non-limiting embodiments of the present invention are described in detail below with reference to the accompanying drawings.

Referring to fig. 2, a method for identifying a credential number area in an image in an embodiment of the present invention is shown, and in this embodiment, the method may include the following steps:

s201: and acquiring an image to be identified.

In this embodiment, the image to be recognized may be obtained from other devices, or may be directly captured by using an image capture module (such as a camera) on the device terminal, where it needs to be emphasized that the image to be recognized may be a picture obtained by picture shooting, or a video frame captured from a video. Wherein the image to be recognized may include an image of a user credential; it should be noted that the image to be identified includes a complete image of the user credential, that is, all images of the user credential are in the image to be identified, and the image of the user credential may be at any position in the image to be identified, and the ratio of the area size of the image of the user credential to the area size of the image to be identified may also be any ratio; that is to say, in this embodiment, it is not necessary for the user to accurately and manually place the user credential at a specific position in the image capturing area, but only the user needs to place the user credential entirely in the image capturing area, so that it can be ensured that the acquired image to be recognized may include all complete images of the user credential, and the area size of the user credential area corresponding to the user credential and the position in the image to be recognized may be arbitrary.

It is emphasized that in this embodiment, a user credential may be understood as a credential that can be used to prove identity information of a user in some aspect, for example, an identity card, a contact/contactless chip card, a magnetic stripe card, a passbook, and the like. It should be noted that, in general, the user credential may include a credential number, and the credential number may be a string of characters; the characters in the string corresponding to the voucher number may all be numeric characters, such as the voucher number may be 6200042535683564; or the partial characters can be numeric characters and the partial characters can be English characters, for example, the certificate number can be date12/26, H620045445; or all English characters, for example, the certificate number may be date.

Next, an example will be described with reference to fig. 3 a. Suppose that fig. 3a is the acquired image to be identified, the user certificate is a bank card, and the certificate number is the card number of the bank card; as shown in fig. 3a, the image to be recognized includes a complete image of a bank card, that is, includes a user certificate area corresponding to the bank card, and all character strings corresponding to the certificate number in the bank card are digital characters.

S202: and determining a user certificate area corresponding to the user certificate in the image to be identified.

In this embodiment, after the image to be recognized is obtained, the user credential area corresponding to the user credential may be recognized in the image to be recognized first, where the user credential area is an image area where the user credential is located in the image to be recognized. Then, the user credential area may be obtained by dividing the user credential area.

In a specific implementation manner, image segmentation may be performed on an image to be recognized to obtain a foreground image region of the image to be recognized, and then the foreground image region is used as a user credential region. The foreground image area can be understood as an image area corresponding to the user credential in the image to be recognized.

Specifically, a foreground image region of the image to be recognized may be determined first. For example, in one implementation, the area to be recognized may be separated from the background area by using an onecut algorithm, so as to obtain the foreground area. For another example, in another implementation manner, edge pixel points in the foreground image region may be determined according to pixel values of respective pixel points in the image to be recognized, where the edge pixel points may be understood as pixel points on edges (i.e., respective edges) belonging to the foreground image region; for example, edge pixels in the foreground image region may be determined according to a change in luminance or a change in pixel value of the pixel, for example, if the luminance change between the pixel a and the pixel B is greater than a preset threshold, the pixel a and the pixel B may be regarded as edge pixels in the foreground image region; then, determining each edge of the foreground image area in the image to be identified according to the edge pixel points; then, the image to be recognized may be segmented according to the edges of the foreground image region to obtain the foreground image region of the image to be recognized, that is, an image region surrounded by the edges may be determined according to the edges, where the image region is the foreground image region in the image to be recognized, that is, the user credential region corresponding to the user credential. After determining the foreground image area of the image to be recognized, image segmentation can be performed on the image to be recognized to obtain the foreground image area of the image to be recognized, and the foreground image area is used as a user certificate area.

It should be noted that, in the process of image recognition processing, if the placement direction of the user credential area may be the horizontal direction, the calculation processing process of the credential number area corresponding to the credential number of the user credential may be subsequently determined, which reduces the calculation complexity and improves the accuracy of determining the credential number area. Therefore, in an implementation manner of this embodiment, after step S202, the method may further include:

determining a primary direction of the user credential area;

and rotating the user certificate area according to the angle difference between the main direction and a preset target direction until the main direction of the user certificate area is the preset target direction.

In this embodiment, a center line direction of a foreground image area (i.e., a user credential area) obtained by performing image segmentation processing on the image to be recognized (i.e., a center line direction of the user credential area in a preset target direction) may be referred to as a main direction of the user credential area. The preset target direction may be preset by a user or preset by a system, and in one possible embodiment, the preset target direction may be a horizontal direction.

Next, an example will be described with reference to fig. 3a, 3b, and 4. Assuming that the preset target direction is a horizontal direction, the user certificate is a bank card, and fig. 4 is a foreground image area obtained by performing image segmentation processing on the image to be identified (i.e., fig. 3a), i.e., a user certificate area corresponding to the bank card; after obtaining the user credential area corresponding to the bank card shown in fig. 4, it may be determined that the main direction of the user credential area is the direction of the center line B of the user credential area; then, the user credential area may be rotated according to the angle difference between the main direction and a preset target direction (i.e., horizontal line a), that is, the user credential area is rotated by the angle difference until the main direction of the user credential area is the preset target direction, that is, the horizontal line a coincides with the central line B, and at this time, the user credential area shown in fig. 3B may be obtained.

S203: and determining a certificate number area corresponding to the certificate number of the user certificate based on the user certificate area.

After determining the user credential area, an area characteristic of the credential number may be identified in the user credential area based on the user credential area and an area characteristic of the credential number area. The region characteristics of the certificate number region can be understood as a certificate number typesetting structure. For example, in one implementation, the character string features of the voucher number may include a character string length, an arrangement manner of the character string, and a character size of each character in the character string, and the character string length, the arrangement manner of the character string, and the character size of each character of the voucher number may be predetermined, for example, when the voucher number is a bank card number, the number of numeric characters of the card number of each bank card of the xx bank is 16 digits, the length is 5cm, the stroke width is 3mm, and the arrangement manner is horizontally arranged. Therefore, the area characteristic of the certificate number area, namely the certificate number typesetting structure, can be determined in advance according to the character string characteristic of the certificate number of the user certificate. It should be noted that, in an implementation manner, when the arrangement manner of the character strings in the credential number is a row or a column, the credential number layout structure may include at least one strip-shaped area, and the width of the strip-shaped area meets a preset credential number area width condition. It should be emphasized that, since the stroke widths of the parts of the characters in the voucher number are all the same under normal conditions, the preset voucher number area width condition may be set in advance according to the stroke widths of the parts of the characters in the voucher number, for example, the preset voucher number area width condition may be that the error between the width of the strip area and the preset width is smaller than a preset threshold, where the preset width may be determined according to the number of words in the voucher number and the width of the characters, for example, may be the product of the number of words in the voucher number and the width of the characters. In addition, it should be emphasized that, since the stroke widths of the respective parts of the characters in the voucher number are all the same, the height and width of each character in the voucher number can be the same, so that, when the voucher number is a row of character strings arranged horizontally, the width of the voucher number area corresponding to the voucher number can be determined according to the number of characters and the character width of the voucher number, and the height of the voucher number area is the height of one character, that is, the width and height of the voucher number area can be fixed and preset.

As an example, one implementation of S203 may include the following steps:

s203 a: and carrying out edge detection on the user certificate area to obtain an edge detection image.

In this embodiment, the edge detection is to detect the edge of the certificate number in the user certificate area, for example, the edge of the bank card number. Specifically, the edge detection method may be: determining a pixel belonging to an edge through the change of the brightness of the pixel in the user certificate area, for example, if the pixel a and the pixel B are two adjacent pixels, if the brightness change between the pixel a and the pixel B is large, the pixel a can be regarded as the pixel belonging to the edge; then, the edge of the voucher number in the user voucher area can be determined according to the pixels belonging to the edge. It will be appreciated that the area bounded by the edges in the user credential area may generally reflect the credential number in the user credential area, such as the card number in a bank card. Therefore, the edge detection image obtained by edge detection of the user certificate area can greatly reduce the data volume in the user certificate area, and can eliminate unimportant information in the user certificate area and reserve important information related to the certificate number in the user certificate area.

Therefore, in this embodiment, after the terminal device obtains the user credential area, the terminal credential may perform edge detection on the user credential area to obtain a detection result of the user credential area, where the detection result may include an edge existing in the user credential area. For example, assuming that the image shown in fig. 3b is a user credential area, after the user credential area is edge-detected, an edge-detected image shown in fig. 3d can be obtained, where the edge-detected image includes edges of a plurality of areas in the user credential area.

It should be noted that the present embodiment provides various methods for performing edge detection on the user credential area, for example, a canny edge detection method or a Sobel (Sobel) edge detection method may be adopted. Next, taking canny edge detection as an example, how to perform edge detection on the user credential area to obtain a detection result will be specifically described.

As an example, since the brightness characteristic in the picture is more obvious than the color gradient characteristic, the accuracy and the speed of identification can be improved by using the processing mode of brightness integration to replace the traditional determination mode of the certificate number area to judge the certificate such as the certificate number (for example, a bank card) independently. Therefore, the user certificate area can be subjected to image graying processing to obtain a grayscale image. It should be noted that, because it is only necessary to determine that the user credential area includes the credential number area carrying the credential number, and the criterion for determining whether one of the user credential areas is the credential number area is not related to the color of the area, that is, the embodiment does not care about the color included in the user credential area. Therefore, in order to improve the accuracy of determining the character area in the image to be recognized, the user certificate area may be subjected to image graying to obtain a grayscale image. Specifically, after the user certificate area is obtained, the image graying process may be performed on the user certificate area, for example, the graying process may be performed on the user certificate area by using a component method, a maximum value method, an average value method, or a weighted average method, so as to obtain a graying result (i.e., a grayscale image) of the user certificate area. It is understood that the graying result of the image to be recognized may include the grayscale value of each pixel in the user credential area. It is emphasized that the gray-scale image of the user-credential area is compared with the user-credential, which only preserves the gray-scale values of the user-credential area, and thus, the gray-scale image enables a dimension reduction of the image to be recognized. Because each pixel in the gray image only has a gray value, only the gray value of each pixel needs to be processed in the process of performing marginalization detection on the gray image, and thus, compared with the process of directly performing marginalization detection on a user certificate area, the calculation amount of marginalization detection on the image to be identified can be reduced. Therefore, the user certificate area under the gray scale result can be subjected to dimension reduction, and only the gray values of all pixels in the user certificate area are reserved, so that the complexity of identifying the certificate number area in the user certificate area is greatly reduced, the identification error caused by redundant information is reduced, and the precision of identifying the certificate number area in the user certificate area is improved.

Then, a brightness mean value can be determined according to the brightness value of each pixel point in the gray level image, and an edge detection image corresponding to the user certificate area is determined according to the brightness mean value. It should be noted that, in general, the brightness of the certificate number in the user certificate is higher than the brightness of other areas in the user certificate, for example, when the user certificate is a bank card, the certificate number, i.e., the card number, in the bank card is usually set in the form of an embossment, and compared with the areas of the bank card except for the card number, the brightness of the card number part is higher and the stroke width of the embossment card number is consistent; therefore, after the gray level image is obtained, the edge detection image corresponding to the user certificate area can be determined according to the brightness value of each pixel point in the gray level image.

Specifically, the mean value of the brightness of all the pixels can be determined according to the brightness value of each pixel in the gray level image. And then, determining an edge detection image corresponding to the user certificate area by using the brightness mean value and a canny edge detection mode. For example, the way of determining the edge detection image by canny edge detection may be: step 1, smoothing the gray level image by using a Gaussian filter to obtain a smooth image, thereby realizing the function of filtering noise of the gray level image; step 2, respectively calculating the gradient strength and the direction of each pixel point in the smooth image; step 3, based on the smooth image, applying Non-Maximum Suppression (Non-Maximum Suppression) to eliminate stray response caused by edge detection; step 4, based on the smooth image obtained in the step 3, determining real and potential edges by using the brightness mean value as Double-Threshold (Double-Threshold) detection; and 5, based on the smooth image obtained in the step 4, finally finishing edge detection by restraining the isolated weak edge. In this way, the edge detection image corresponding to the user credential area can be determined.

S203 b: and performing edge expansion processing on the user certificate area according to the edge detection image to obtain an edge expansion image.

In this embodiment, in the process of determining the voucher number area, only the voucher number in the image to be identified needs to be determined, and in the edge detection image obtained by performing edge detection, burrs may exist at the edge of the area, and the burrs may affect the judgment on whether the area is the voucher number area; in addition, since the image to be recognized may be an image of light received from the shooting environment during shooting, there is a problem that the light interferes.

Therefore, after the user certificate area is subjected to edge detection to obtain an edge detection image, edge expansion processing can be performed on the user certificate area according to the edge detection image to obtain an edge expansion image. For example, the edge detection image can be processed by Gaussian blur filtering to obtain an edge expansion image with an expanded edge range, so that interference factors of shooting field light can be reduced, burrs in the edge can be eliminated, the edge expansion image is free of burrs, the outline of the certificate number area is more obvious, and the accuracy of determining the certificate number area according to the edge expansion image can be guaranteed.

It should be noted that the edge expansion is understood as a process of merging all background pixels in contact with the edge in the user credential area into the edge and expanding the edge to the outside to some extent.

S203 c: and if the edge expansion image has an image area which accords with the certificate number typesetting structure, taking the image area which accords with the certificate number typesetting structure as the certificate number area, and intercepting the certificate number area.

In this embodiment, after obtaining the edge expanded image, it may be determined whether there is an image area conforming to the voucher number layout structure in the edge expanded image according to the voucher number layout structure, and if so, the image area conforming to the voucher number layout structure may be used as the voucher number area, and the voucher number area may be intercepted. For example, if the voucher number typesetting structure includes at least one strip-shaped area, and the width of the strip-shaped area meets the preset voucher number area width condition, it may be determined whether the strip-shaped area exists in the edge expanded image, and if so, it may be further determined whether the strip-shaped area meets the preset voucher number area width condition; if the strip-shaped area meets the width condition of the preset certificate number area, the strip-shaped area can be used as the certificate number area, and the certificate number area is intercepted. For example, referring to fig. 3g, assuming that the user certificate area is the image shown in fig. 3b, after the edge expanded image corresponding to the user certificate area is obtained, since there is an image area (as shown in fig. 3 g) conforming to the certificate number layout structure in the edge expanded image, the image area conforming to the certificate number layout structure may be used as the certificate number area, and the certificate number area is intercepted, so as to obtain the certificate number area shown in fig. 3 g.

It should be noted that, in the technical solution of this embodiment, it is determined whether the image to be recognized includes the credential number area carrying the credential number according to the credential number typesetting structure, that is, this embodiment may determine the credential number area carrying the credential number by the outline of each area in the user credential area. Therefore, in order to improve the accuracy and efficiency of determining the voucher number area in the image to be identified, in an implementation manner of this embodiment, after the step of performing edge dilation processing on the user voucher area according to the edge detection image to obtain an edge dilated image, the method may further include:

and carrying out binarization processing on the edge expansion image to obtain a binarization image.

Specifically, after the edge extension processing is performed on the user credential area to obtain an edge extension image, binarization processing may be performed on an edge in the edge extension image. For example, in the edge-dilated image, all pixels with gray values greater than or equal to the threshold value in the edge of the edge-dilated image may be regarded as belonging to the edge, and the gray values of the pixels with gray values greater than or equal to the threshold value may be set to 255; and all pixels of which the gray values are smaller than the threshold value in the edge detected by the detection result can be regarded as not belonging to the edge, and the gray values of the pixels of which the gray values are smaller than the threshold value can be set to be 0 to represent the background. Next, as an example, with reference to fig. 3e, assuming that the user certificate area is the image shown in fig. 3b, after an edge extension image corresponding to the user certificate area is obtained, binary processing is performed on the edge extension image to obtain a binary image (certificate number area shown in fig. 3 e).

After the edge expansion image is subjected to binarization processing, a binarized image of the edge expansion image with the edge in the edge expansion image can be obtained. In the process of performing binarization processing on the edge expansion image, the gray value of each pixel of the edge in the edge expansion image to be identified in the edge expansion image can be set to be 0 or 255. Therefore, the binary image can show obvious black and white effect, and compared with the edge expansion image, the data volume of the binary image is greatly reduced, and the outline of each edge in the user certificate area can be more highlighted.

Correspondingly, if the image area conforming to the voucher number typesetting structure exists in the edge expanded image, taking the image area conforming to the voucher number typesetting structure as the voucher number area, and intercepting the voucher number area, including:

and if the image area which accords with the certificate number typesetting structure exists in the binary image, taking the image area which accords with the certificate number typesetting structure as the certificate number area, and intercepting the certificate number area.

As an example, the way to determine the voucher number area from the binarized image may be:

and firstly determining one-dimensional brightness statistic values corresponding to a plurality of target areas in the binary image. The determination method of the one-dimensional luminance statistic corresponding to each target area is as follows: determining a target area by taking pixel points of which the pixel values meet a preset value (for example, the pixel values are 255) in the binarized image as central pixel points and taking the stroke width of the preset certificate number as a width range, for example, determining the target area corresponding to the central pixel points by taking the pixel points of which the pixel values are 255 in the binarized image as the central pixel points and taking the central pixel points as the centers and the stroke width of the preset certificate number as the width range; then, the gray values of all the pixel points in the target region may be summed, and the central pixel point is marked with the sum value, so as to obtain an edge brightness statistical image (it should be noted that the edge brightness statistical image may be determined according to all the central pixel points, for example, fig. 3f is the edge brightness statistical image corresponding to fig. 3 b); obtaining a one-dimensional brightness value sequence in the vertical direction according to the product value in the vertical direction in the edge brightness statistical image, for example, obtaining a one-dimensional brightness value sequence in the vertical direction of the center pixel point according to the product value in the vertical direction of the center pixel point in the edge brightness statistical image; and calculating a one-dimensional luminance statistic value on the one-dimensional luminance value sequence by taking the preset height of the voucher number area as a span, for example, taking the central pixel point as a starting point, an end point or a middle point, and taking the preset height of the voucher number area as a span, determining all pixel points in the span, and determining the one-dimensional luminance statistic value on the one-dimensional luminance value sequence of the central pixel point according to the luminance of all the pixel points in the span.

After the target area corresponding to the pixel point of which the pixel value in each binary image meets the preset value is obtained, the position of the pixel point corresponding to the maximum value in the one-dimensional luminance statistics values corresponding to the plurality of target areas can be used as the initial position of the certificate number in the certificate number area in the vertical direction. Then, the voucher number area can be determined according to the starting position, and the voucher number area is intercepted. For example, since the height and width of the voucher number area can be determined in advance according to the character size and the number of characters of the character string of the voucher number in the user voucher, the position of the pixel point corresponding to the maximum value in the one-dimensional luminance statistics corresponding to the plurality of target areas can be used as the starting point after the voucher number is located at the starting position in the vertical direction in the voucher number area (i.e., the position of the pixel point corresponding to the maximum value in the one-dimensional luminance statistics corresponding to the plurality of target areas), and the height and width of the preset voucher number area are combined to determine the voucher number area, and the voucher number area is intercepted.

It should be noted that, after determining the credential number area corresponding to the credential number in the user credential area, the credential number in the credential number area may be further determined, so as to facilitate subsequent processing (e.g., inputting the credential number) on the credential number. In an implementation manner of this embodiment, after S203, the embodiment corresponding to fig. 2 may further include the following steps:

and performing text recognition processing on the certificate number area to obtain the certificate number.

In this embodiment, the document number area may be subjected to text recognition processing by using a conventional image recognition algorithm, or may be subjected to text recognition processing by using a neural network, so as to obtain a document number. In this way, in the embodiment, the voucher number area is determined in the image to be identified, and then the text identification processing is performed on the voucher number area, that is, other image areas irrelevant to the voucher number in the image to be identified are filtered out, and then the voucher number area with higher relevance to the voucher number is identified; compared with the image to be identified, the data volume in the image is greatly reduced in the voucher number area, so that the technical scheme provided by the application has small calculation amount compared with the traditional image identification algorithm processing mode, the voucher number in the image to be identified can be rapidly and accurately determined, the requirement of identifying a large number of images can be met, and the identification cost is reduced.

Next, a method of performing text recognition processing on the voucher number area to obtain the voucher number will be described, taking a method of performing text recognition processing on the voucher number area to obtain the voucher number as an example.

Specifically, the credential number area may be input to a trained number recognition model, and the credential number output by the number recognition model is obtained. The number recognition model may be obtained by training based on a training image set, the training image set includes a plurality of training image subsets, each training image subset may include a training image at least including a string of character strings and character strings corresponding to the training images, all character strings in the character strings may be digital characters, or all character strings may be english characters, or some characters may be digital character strings, and some characters may be english characters. It is emphasized that, when the user certificate is a bank card, the card number of the bank card is digital characters, so the training image set includes a plurality of training image subsets, each training image subset may include a training image including at least one digital character and a digital character string corresponding to the training image; it is emphasized that the training image set at least includes digital characters corresponding to ten digits, i.e., the digits 0 to 9, and in one implementation, the training images in the training image set may each be an image including digital characters in an embossed font, or an image including digital characters in a non-embossed font, or an image in which the digital characters in the embossed font and the digital characters in the non-embossed font are mixed; in this way, since the trained image content only includes ten numeric characters, the requirement on the complexity of the network structure of the neural network can be reduced, so that the training image set can be trained by using a simplified network model, the training speed is increased, and the identification speed and the identification accuracy of the voucher number can also be increased.

Therefore, the mode of performing text recognition processing on the certificate number area by using the neural network to obtain the certificate number can provide convenience for subsequent processing (for example, inputting the certificate number) on the certificate number.

It should be noted that, in this embodiment, the above-mentioned algorithm part may be completed by calling an algorithm dynamic library using a python code, and since the python code is time-consuming when executing a loop operation, the technical solution uses cpython to accelerate the loop part in the algorithm, so that the overall recognition speed is obviously improved. Python is a script type parsing execution, parsing operation of a loop part is slow in execution along with accumulation of loop times, cpython can compile pyx (with slightly changed Python syntax) for execution, and a loop calculation part is realized and compiled by using pyx for Python calling.

Therefore, the embodiment realizes the voucher number area identification method process in the image by combining with a specific application scene. Of course, the above scenario is only an exemplary scenario and is not intended to limit the method provided by the present invention. The method provided by the invention can be applied to the identification process of the certificate number area in other images with the same principle.

Fig. 5 shows a specific embodiment of the document number area recognition device in the image according to the present invention. The apparatus of this embodiment is a physical apparatus for executing the method of the above embodiment. The technical solution is essentially the same as the above embodiments, and the corresponding descriptions in the above embodiments are also applicable to this embodiment. The device in this embodiment includes:

an image obtaining unit 501, configured to obtain an image to be identified, where the image to be identified includes an image of a user credential;

a first determining unit 502, configured to determine a user credential area corresponding to the user credential in the image to be identified;

a second determining unit 503, configured to determine, based on the user credential area, a credential number area corresponding to a credential number of the user credential.

Optionally, the first determining unit 502 is specifically configured to:

carrying out image segmentation processing on the image to be identified to obtain a foreground image area of the image to be identified; the foreground image area is an image area corresponding to the user certificate in the image to be identified;

and taking the foreground area as the user certificate area.

Optionally, the first determining unit 502 is specifically configured to:

determining edge pixel points of a foreground image area according to pixel values of all pixel points in the image to be identified;

determining each edge of a foreground image area in the image to be identified according to the edge pixel points;

and segmenting the image to be identified according to each edge of the foreground image area to obtain the foreground image area of the image to be identified.

Optionally, the apparatus further includes a third determining unit, configured to:

determining a primary direction of the user credential area;

Optionally, the second determining unit 503 is specifically configured to:

performing edge detection on the user certificate area to obtain an edge detection image;

performing edge expansion processing on the user certificate area according to the edge detection image to obtain an edge expansion image;

if the edge expansion image has an image area which accords with the certificate number typesetting structure, taking the image area which accords with the certificate number typesetting structure as the certificate number area, and intercepting the certificate number area.

Optionally, the second determining unit 503 is specifically configured to:

carrying out image graying processing on the user certificate area to obtain a grayscale image;

determining a brightness mean value according to the brightness value of each pixel point in the gray level image;

and determining an edge detection image corresponding to the user certificate area according to the brightness mean value.

Optionally, the voucher number typesetting structure includes at least one strip-shaped area, and the width of the strip-shaped area meets the preset voucher number area width condition.

Optionally, the apparatus further includes a binarization unit, configured to:

carrying out binarization processing on the edge expansion image to obtain a binarization image;

correspondingly, the second determining unit 503 is specifically configured to:

Optionally, the second determining unit 503 is specifically configured to:

determining one-dimensional brightness statistic values corresponding to a plurality of target areas in the binary image; the determination method of the one-dimensional luminance statistic corresponding to each target area is as follows: determining a target area by taking pixel points with pixel values meeting a preset value in the binary image as central pixel points and taking the stroke width of the preset certificate number as a width range; summing the gray values of all pixel points in the target area, marking the central pixel point with the sum value, and obtaining an edge brightness statistical image; obtaining a one-dimensional brightness value sequence in the vertical direction according to the product value in the vertical direction in the edge brightness statistical image; calculating a one-dimensional brightness statistic value with the preset height of the certificate number area as a span on the one-dimensional brightness value sequence;

taking the position of a pixel point corresponding to the maximum value in the one-dimensional brightness statistics values corresponding to the target areas as the initial position of the certificate number in the certificate number area in the vertical direction;

and determining the certificate number area according to the starting position, and intercepting the certificate number area.

Optionally, the apparatus further includes a text recognition unit, configured to:

Optionally, the text recognition unit is specifically configured to:

inputting the certificate number area into a trained number recognition model to obtain the certificate number output by the number recognition model;

the number recognition model is obtained by training based on a training image set, the training image set comprises a plurality of groups of training image subsets, each group of training image subsets comprises a training image at least comprising one digital character and a digital character string corresponding to the training image. .

Fig. 6 is a schematic structural diagram of an electronic device according to an embodiment of the present invention. On the hardware level, the electronic device comprises a processor and optionally an internal bus, a network interface and a memory. The Memory may include a Memory, such as a Random-Access Memory (RAM), and may further include a non-volatile Memory, such as at least 1 disk Memory. Of course, the electronic device may also include hardware required for other services.

The processor, the network interface, and the memory may be connected to each other via an internal bus, which may be an ISA (Industry Standard Architecture) bus, a PCI (peripheral component Interconnect) bus, an EISA (Extended Industry Standard Architecture) bus, or the like. The bus may be divided into an address bus, a data bus, a control bus, etc. For ease of illustration, only one double-headed arrow is shown in FIG. 6, but that does not indicate only one bus or one type of bus.

And the memory is used for storing the execution instruction. In particular, a computer program that can be executed by executing instructions. The memory may include both memory and non-volatile storage and provides execution instructions and data to the processor.

In a possible implementation manner, the processor reads the corresponding execution instruction from the nonvolatile memory to the memory and then runs the corresponding execution instruction, and the corresponding execution instruction can also be obtained from other equipment so as to form the certificate number area identification device in the image on a logic level. The processor executes the execution instruction stored in the memory, so that the certificate number area identification method in the image provided by any embodiment of the invention is realized through the executed execution instruction.

The method executed by the credential number area recognition device in the image provided by the embodiment of the invention shown in fig. 2 can be applied to or implemented by a processor. The processor may be an integrated circuit chip having signal processing capabilities. In implementation, the steps of the above method may be performed by integrated logic circuits of hardware in a processor or instructions in the form of software. The Processor may be a general-purpose Processor, including a Central Processing Unit (CPU), a Network Processor (NP), and the like; but also Digital Signal Processors (DSPs), Application Specific Integrated Circuits (ASICs), Field Programmable Gate Arrays (FPGAs) or other Programmable logic devices, discrete Gate or transistor logic devices, discrete hardware components. The various methods, steps and logic blocks disclosed in the embodiments of the present invention may be implemented or performed. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.

The steps of the method disclosed in connection with the embodiments of the present invention may be directly implemented by a hardware decoding processor, or implemented by a combination of hardware and software modules in the decoding processor. The software module may be located in ram, flash memory, rom, prom, or eeprom, registers, etc. The storage medium is located in a memory, and a processor reads information in the memory and completes the steps of the method in combination with hardware of the processor.

The embodiment of the present invention further provides a readable storage medium, where the readable storage medium stores an execution instruction, and when the stored execution instruction is executed by a processor of an electronic device, the electronic device can execute the method for identifying the credential number area in the image provided in any embodiment of the present invention, and is specifically configured to execute the method for identifying the credential number area in the image.

The electronic device described in the foregoing embodiments may be a computer.

It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as a method or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects.

The embodiments of the present invention are described in a progressive manner, and the same and similar parts among the embodiments can be referred to each other, and each embodiment focuses on the differences from the other embodiments. In particular, as for the apparatus embodiment, since it is substantially similar to the method embodiment, the description is relatively simple, and for the relevant points, reference may be made to the partial description of the method embodiment.

It should also be noted that the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises the element.

The above description is only an example of the present invention, and is not intended to limit the present invention. Various modifications and alterations to this invention will become apparent to those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the scope of the claims of the present invention.

Claims

1. A method for identifying a voucher number area in an image, the method comprising:

2. The method according to claim 1, wherein the determining a user credential area corresponding to the user credential in the image to be recognized comprises:

and taking the foreground area as the user certificate area.

3. The method according to claim 2, wherein the performing image segmentation processing on the image to be recognized to obtain a foreground image region of the image to be recognized comprises:

determining edge pixel points of a foreground image area according to the pixel values of all the pixel points in the image to be identified;

4. The method according to any of claims 1-3, wherein after the step of determining a user credential area to which the user credential corresponds in the image to be recognized, the method further comprises:

determining a primary direction of the user credential area;

5. The method of claim 1, wherein the determining, based on the user credential area, a credential number area corresponding to a credential number of the user credential comprises:

and if the edge expansion image has an image area which accords with the certificate number typesetting structure, taking the image area which accords with the certificate number typesetting structure as the certificate number area, and intercepting the certificate number area.

6. The method of claim 5, wherein the edge detecting the user credential area to obtain an edge detection image comprises:

7. The method according to claim 6, wherein the voucher number layout structure comprises at least one strip-shaped area, and the width of the strip-shaped area meets a preset voucher number area width condition.

8. The method according to claim 7, wherein after the step of performing edge dilation processing on the user credential area according to the edge detection image to obtain an edge dilation image, the method further comprises:

9. The method according to claim 8, wherein if the binarized image has an image area conforming to the document number layout structure, taking the image area conforming to the document number layout structure as the document number area, and intercepting the document number area, comprises:

10. The method of claim 1, wherein after the step of determining a credential number region corresponding to a credential number of the user credential based on the user credential region, the method further comprises:

11. The method of claim 10, wherein the text recognition processing of the voucher number area to obtain the voucher number comprises:

the number recognition model is obtained by training based on a training image set, the training image set comprises a plurality of groups of training image subsets, each group of training image subsets comprises a training image at least comprising one digital character and a digital character string corresponding to the training image.

12. An apparatus for identifying a voucher number area in an image, the apparatus comprising:

the device comprises an image acquisition unit, a recognition unit and a recognition unit, wherein the image acquisition unit is used for acquiring an image to be recognized, and the image to be recognized comprises an image of a user certificate;

13. An electronic device comprising a processor and a memory storing execution instructions, the processor performing the method of any one of claims 1-11 when the processor executes the execution instructions stored by the memory.