CN111144400B - Identification method, device, terminal equipment and storage medium for ID card information - Google Patents

Abstract

The invention provides an identification method, a device, terminal equipment and a storage medium of identity card information, and relates to the technical field of intelligent identification, wherein the method comprises the following steps: acquiring an identity card image of identity card information to be identified; performing text line detection on the identity card image to obtain a plurality of text boxes; acquiring position information of a text box in an identity card image; performing text recognition on the text box to obtain text information corresponding to the text box; determining text information with a corresponding relation according to the position information and the text information of the text box; and determining the identity card information corresponding to the identity card image according to the text information with the corresponding relation. The invention can effectively improve the identification accuracy of the identity card information.

Description

Identification method, device, terminal equipment and storage medium for ID card information

Technical field

The present invention relates to the field of intelligent identification technology, and in particular to an identification method, device, terminal equipment and storage medium for ID card information.

Background technique

At present, many occasions require identification of ID card information to determine the identity of the ID card holder. In the existing technology, most of the identity information on the ID card is directly read through hardware devices such as ID card readers.

With the development of technology, ID card image recognition technology has gradually emerged. It mainly recognizes and extracts text from ID card images, thereby extracting ID card information, such as name, gender, date of birth, ID card number, address, etc. However, due to the large amount of text information on the ID card, the extracted text information is relatively scattered, other text such as watermarks may be added to the ID card image, and the ID card image is likely to be tilted. Therefore, it is possible to extract wrong ID card information, and the recognition accuracy of ID card information is not high. For example, it is possible to obtain an incorrect information correspondence when performing text recognition. For example, when the ID card image is tilted, "name" and "male" may be considered as a group, thereby drawing the wrong conclusion that the name is "male".

Contents of the invention

In view of this, the purpose of the present invention is to provide a method, device, terminal equipment and storage medium for identifying ID card information, which can effectively improve the identification accuracy of ID card information.

In a first aspect, embodiments of the present invention provide a method for identifying ID card information, including:

Obtain an ID card image of the ID card information to be identified; perform text line detection on the ID card image to obtain multiple text boxes; obtain position information of the text boxes in the ID card image; perform text recognition on the text boxes to obtain text information corresponding to the text boxes; determine text information with a corresponding relationship based on the position information and text information of the text boxes; determine the ID card information corresponding to the ID card image based on the text information with a corresponding relationship.

In conjunction with the first aspect, an embodiment of the present invention provides a first possible implementation manner of the first aspect, wherein the above-mentioned step of detecting text lines on the ID card image to obtain multiple text boxes includes: obtaining the ID card image Based on the vertex coordinates of the ID card image, determine whether the ID card image is tilted; if so, correct the ID card image; perform text line detection on the adjusted ID card image to obtain multiple text boxes.

Combined with the first aspect, embodiments of the present invention provide a second possible implementation manner of the first aspect, in which the above-mentioned step of detecting text lines on the ID card image to obtain multiple text boxes includes: using a CTPN network model to The ID card image is subjected to text line detection and multiple text boxes are obtained.

In conjunction with the first aspect, embodiments of the present invention provide a third possible implementation of the first aspect, wherein the above-mentioned step of obtaining the position information of the text box in the ID card image includes: obtaining the vertices of the text box in the ID card image. coordinates; based on the vertex coordinates of the text box, determine whether the text box is tilted; if so, adjust the text box; obtain the position information of the adjusted text box.

Combined with the first aspect, embodiments of the present invention provide a fourth possible implementation manner of the first aspect, wherein the above step of performing text recognition on a text box and obtaining text information corresponding to the text box includes: using a CRNN network model to The text box performs text recognition and obtains the text information corresponding to the text box.

In combination with any one of the first aspect to the fourth possible implementation manner of the first aspect, embodiments of the present invention provide a fifth possible implementation manner of the first aspect, wherein the above-mentioned position information and text information based on the text box , the steps of determining text information with corresponding relationships include: determining text boxes of the same peer based on the position information of the text boxes; determining text boxes with corresponding relationships among the text boxes of the same peers based on the location information of the text boxes of the same peers; The corresponding text boxes and the text information of the text boxes determine the text information with the corresponding relationship.

In conjunction with the fifth possible implementation manner of the first aspect, embodiments of the present invention provide a sixth possible implementation manner of the first aspect, wherein the position information of the text box includes the starting point coordinates of the text box, the starting point coordinates It includes a starting abscissa and a starting ordinate; the step of determining the text boxes of the same row according to the position information of the text box includes: determining the text boxes whose ordinates differ from the starting point within a preset first threshold as the text boxes of the same row; According to the position information of the text boxes of the same peer, the step of determining the text boxes with corresponding relationships in the text boxes of the same peer includes: determining the text boxes of the same peer that are within a preset second threshold of the starting point abscissa difference of the text boxes of the same peer. Text boxes with corresponding relationships within text boxes.

In combination with the five possible implementations of the first aspect, embodiments of the present invention provide a seventh possible implementation of the first aspect, wherein the position information of the text box includes the ordinate value range and the abscissa of the text box. The value range; the step of determining the text boxes of the same row according to the position information of the text boxes includes: determining two text boxes whose ordinate value range has at least the first overlapping length as the text boxes of the same row; according to the text boxes of the same row The steps of determining the corresponding text boxes in the text boxes of the same row include: determining the minimum value of the abscissa coordinate of the text box according to the value range of the abscissa coordinate of the text box of the same row; The text boxes whose minimum values differ within the preset third threshold are determined to be text boxes that have a corresponding relationship among the text boxes of the same row.

In combination with the five possible implementations of the first aspect, embodiments of the present invention provide an eighth possible implementation of the first aspect, wherein, based on the text boxes with corresponding relationships and the text information of the text boxes, it is determined that the The steps of corresponding text information include: checking whether the corresponding relationship of the text box is correct based on the text box with the corresponding relationship and the text information of the text box; determining the text information of the text box with the correct corresponding relationship as having the corresponding relationship text information.

In combination with the first aspect, an embodiment of the present invention provides a ninth possible implementation of the first aspect, wherein the step of determining the identity card information corresponding to the identity card image based on the text information with a corresponding relationship includes: determining the information category of the text information with a corresponding relationship based on the text information with a corresponding relationship; the information category includes one or more of the name category, gender category, ethnicity category, date of birth category, address category and identity card number category; according to the information category, determining the keyword words in the text information with a corresponding relationship and the text with a corresponding relationship with the keyword words; wherein the keyword words include one or more of the name, gender, ethnicity, birth, year, month, day, address and citizen identity card number; determining the text with a corresponding relationship with the keyword words as the identity card information.

In combination with the ninth possible implementation manner of the first aspect, embodiments of the present invention provide a tenth possible implementation manner of the first aspect. The above method further includes: checking whether the ID card information is correct according to the information category; if not, Initiate an error message.

In combination with the tenth possible implementation of the first aspect, an embodiment of the present invention provides an eleventh possible implementation of the first aspect, wherein the above-mentioned step of checking whether the ID card information is correct according to the information category includes: using the information corresponding to the ID card number category in the ID card information to check whether the information corresponding to the date of birth category in the ID card information and the information corresponding to the gender category in the ID card information are correct; and/or, checking whether the information corresponding to the gender category in the ID card information is one of the preset gender words, wherein the preset gender words include male or female; and/or, checking whether the information corresponding to the ethnic category in the ID card information is one of the preset ethnic words, wherein the preset ethnic words include legal ethnic words; and/or, checking whether the information corresponding to the date of birth category in the ID card information is one of the preset numbers; wherein the preset numbers are legal date numbers.

In combination with the tenth possible implementation manner of the first aspect, embodiments of the present invention provide a twelfth possible implementation manner of the first aspect. The above method further includes: judging based on the information corresponding to the ID card number category in the ID card information. Is the ID card image legal?

Combined with the first aspect, embodiments of the present invention provide a thirteenth possible implementation manner of the first aspect, wherein before the step of character recognition of the text box, the above method further includes: based on the position information of the text box and the preset settings The legal position of the ID card text box is determined to determine whether there is a text box with an illegal position; if so, delete the text box with an illegal position.

In conjunction with the first aspect, embodiments of the present invention provide a fourteenth possible implementation manner of the first aspect, wherein before the step of determining text information with a corresponding relationship, the above method further includes: based on the position information of the text box and the text Information, identify and filter illegal text information on the ID card image; among them, the position of the text box where the illegal text information is located does not belong to the legal position of the preset ID card text box.

In a second aspect, an embodiment of the present invention provides an identification device for identity card information, comprising: an image acquisition module for acquiring an identity card image of identity card information to be identified; a text line detection module for performing text line detection on the identity card image to obtain multiple text boxes; a position information acquisition module for acquiring position information of the text box in the identity card image; a text recognition module for performing text recognition on the text box to obtain text information corresponding to the text box; a corresponding relationship determination module for determining text information with a corresponding relationship based on the position information and text information of the text box; and an identity card information determination module for determining the identity card information corresponding to the identity card image based on the text information with a corresponding relationship.

In a third aspect, embodiments of the present invention provide a terminal device. The terminal device includes a memory and a processor. The memory is used to store any of the fourteenth possible implementation modes of the first aspect to the first aspect. In a program of the method, the processor is configured to execute the program stored in the memory.

In a fourth aspect, embodiments of the present invention provide a computer storage medium for storing computer software instructions used for the method described in any one of the first aspect to the fourteenth possible implementation manner of the first aspect.

The invention provides a method, device, terminal equipment and storage medium for identifying ID card information, which can first detect text lines on an ID card image to be recognized to obtain multiple text boxes; and then obtain the text in the ID card image. The position information of the frame and the text information corresponding to the text box, and then based on the position information and text information of the text box, the corresponding text information is determined, and based on the corresponding text information, the ID card information corresponding to the ID card image is determined . Unlike the existing technology, which mostly directly recognizes the text on the ID card image to extract the ID card information, the wrong ID card information may be extracted due to reasons such as scattered text on the ID card image, watermarks added or the ID card image being tilted. Compared with the problem, the embodiment of the present invention can first detect the text box on the ID card image, and then determine the corresponding text information based on the position information and text information of the text box, which can avoid extracting the wrong ID card information and effectively Improved the recognition accuracy of ID card information.

Additional features and advantages of the invention will be set forth in the description which follows, and in part will be apparent from the description, or may be learned by practice of the invention. The objectives and other advantages of the invention may be realized and attained by the structure particularly pointed out in the written description, claims and appended drawings.

In order to make the above-mentioned objects, features and advantages of the present invention more obvious and understandable, preferred embodiments are given below and described in detail with reference to the accompanying drawings.

Description of drawings

In order to more clearly explain the specific embodiments of the present invention or the technical solutions in the prior art, the accompanying drawings that need to be used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings in the following description The drawings illustrate some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without exerting any creative effort.

Figure 1 is a flow chart of the first identification method of ID card information provided by an embodiment of the present invention;

Figure 2 is a schematic diagram of a tilted ID card provided by an embodiment of the present invention;

Figure 3 is a schematic diagram of an adjusted ID card provided by an embodiment of the present invention;

Figure 4 is a flow chart of a second identification method for ID card information provided by an embodiment of the present invention;

5 is a flow chart of a third method for identifying identity card information provided by an embodiment of the present invention;

Figure 6 is a structural block diagram of a first identification device for ID card information provided by an embodiment of the present invention;

Figure 7 is a structural block diagram of a second identification device for ID card information provided by an embodiment of the present invention;

Figure 8 is a structural block diagram of a third identification device for ID card information provided by an embodiment of the present invention;

FIG. 9 is a schematic diagram of the structure of a terminal device provided in an embodiment of the present invention.

Detailed ways

In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of them. Embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the scope of protection of the present invention.

Considering that most of the existing technologies directly perform text recognition and extraction on ID card images, the recognition accuracy is not high. In order to improve this problem, embodiments of the present invention provide a method, device, terminal equipment, and device for identifying ID card information. The storage medium can first monitor the text box on the ID card image, and then determine the corresponding text information based on the position information and text information of the text box to avoid extracting the wrong ID card information. This method can effectively improve Recognition accuracy of ID card information. This technology can be applied to any occasion where identification of ID card images is required.

The embodiments of the present invention are introduced in detail below.

The embodiment of the present invention first provides a first method for identifying ID card information, as shown in Figure 1. This method can be executed by a terminal device with processing capabilities such as a computer, a mobile phone, a handheld ID recognition terminal, etc. The method specifically includes the following step:

Step S102, obtaining an ID card image of the ID card information to be identified; wherein the ID card image to be identified can be an electronic ID card print uploaded directly by humans, or an ID card image obtained by directly scanning a physical ID card through an electronic device, etc., and the method for obtaining the ID card image is not limited here. Step S104, performing text line detection on the ID card image to obtain multiple text boxes.

During specific implementation, text detection can be performed on the ID card image through the relevant text line detection algorithm, and the shape of the resulting text box is usually a rectangle.

During specific implementation, a text box detection model can be constructed through machine learning. The input of the model is an ID card image, and the output is each text box on the ID card image.

Step S106: Obtain the position information of the text box in the ID card image.

In an optional implementation, a plane rectangular coordinate system can be constructed on a plane where multiple text boxes coexist, and the coordinates of a certain point at the same position in each text box can be used to represent the position information of its corresponding text box. For example, use the vertex coordinates of the upper left corner of each text box as the position information of the text box. In addition, you can also select multiple key points of each text box and determine the coordinate set of multiple key points as the position information of each text box. Among them, the key point can be a point on the border of the text box. The number of key points can be set according to actual needs and is not limited here.

Step S108: Perform text recognition on the text box to obtain text information corresponding to the text box.

During specific implementation, text recognition of the text box can be performed through relevant text recognition algorithms to obtain text information corresponding to the text box.

During specific implementation, a text box text recognition model can be constructed through machine learning. The input of the model is the image corresponding to the text box in the ID card image, and the output is the text information in the text box image.

Step S110: Determine text information with a corresponding relationship based on the position information and text information of the text box.

For example, according to the position distribution characteristics of the text boxes on the ID card that actually have text correspondences, such as "name" and "Zhang San", "gender" and "female", "ethnicity" and "Han" are far apart in position distribution Recently, the position information of the text box obtained in step S106 was analyzed and judged based on the position distribution characteristics of the corresponding text box, and it was found that "gender" and "female" have a corresponding relationship, and "ethnicity" and "Han" have a corresponding relationship. , that is, the text information with corresponding relationships is initially determined. Considering that the text boxes where "female" and "ethnicity" are located may be close to each other, it is possible to draw the wrong conclusion that "female" and "ethnicity" correspond to each other. Therefore, this embodiment further combines the text information to verify the two relative Whether there is a correspondence between adjacent text boxes, so as to eliminate text boxes with similar position distribution but no text correspondence, such as "female" and "ethnic", etc., and filter out the text boxes that both conform to the position distribution characteristics and have Text boxes for text correspondence, thereby determining corresponding text information. During actual execution, the order of the above-mentioned steps S106 and S108 may not be distinguished.

For example, under normal circumstances, the text information corresponding to the text information "female" and "male" is "gender". When it is determined based on the positional relationship between the text boxes, the text box where the text information "female" is located and the text information "ethnicity" are located. When the text box has a corresponding relationship, further based on the text information "female" and "ethnicity", the text information that should correspond to "female" is "gender" rather than "ethnicity". Therefore, it can be determined that "female" and "ethnicity" are not the same. correspond.

Step S112: Determine the ID card information corresponding to the ID card image based on the text information with the corresponding relationship.

After knowing the corresponding text information, the ID card information can be determined. The ID card information may include at least one of name, gender, date of birth, address, and ID number.

For example, according to the above step S110, it is determined that the "name" in the ID card image has a corresponding relationship with "Zhang San", the "gender" has a corresponding relationship with "female", and the "nationality" has a corresponding relationship with "Han", and the determined ID card information can be, name: Zhang San, gender: female, nationality: Han. The determined ID card information can be presented, specifically, the ID card information can be presented as "name-Zhang San gender-female nationality-Han", or can be directly presented as "Zhang San-female-Han", and the presentation form of the determined ID card information can be set according to needs, and is not limited here.

The above-mentioned identification method of ID card information provided by the present invention can first perform text line detection on the ID card image to be recognized to obtain multiple text boxes; and then obtain the position information of the text boxes in the ID card image, and the corresponding text boxes. text information, and then determine corresponding text information based on the position information and text information of the text box, and determine ID card information corresponding to the ID card image based on the corresponding text information. Unlike the existing technology, which mostly directly recognizes the text on the ID card image to extract the ID card information, there is a possibility that the wrong ID card information is extracted due to reasons such as scattered text on the ID card image, watermarks added or the ID card image being tilted. Compared with the problem, the embodiment of the present invention can first detect the text box on the ID card image, and then determine the corresponding text information based on the position information and text information of the text box, which can avoid extracting the wrong ID card information and effectively Improved the recognition accuracy of ID card information.

In an optional implementation, the above step S104, that is, the step of detecting text lines on the ID card image to obtain multiple text boxes, includes:

(1) Obtain the vertex coordinates of the ID card image, and determine whether the ID card image is tilted based on the vertex coordinates of the ID card image. Among them, the vertex coordinates of the ID card image can be the upper left corner vertex, the lower left corner vertex, the upper right corner vertex and the lower right corner vertex on the ID card image. When judging whether the ID card image is tilted based on the vertex coordinates, you can determine whether the abscissas of the upper left vertex and the lower left vertex of the ID card image are the same, and whether the ordinates of the upper left vertex and the upper right vertex are the same. If they are different, then Explain that the ID card image is tilted.

(2) When the judgment result in (1) is yes, adjust the ID card image.

Among them, the abscissa coordinate of the upper left vertex of the corrected ID card image is the same as the abscissa of the lower left vertex, and the ordinate of the upper left vertex is the same as the ordinate of the upper right vertex, and the distance between the upper left vertex and the upper right vertex is greater than The distance between the upper left vertex and the lower left vertex.

The adjustment method is not limited and can be reasonably selected by those skilled in the art.

(3) Perform text line detection on the corrected ID card image to obtain multiple text boxes.

In another optional implementation, when performing the above step S104, the ID card image may not be adjusted, but text lines may be directly detected on the ID card image. If the detected text box is skewed, the text may be directly Adjust the position of the frame.

Regardless of the implementation method, the CTPN network model can be used to detect text lines on the ID card image to obtain multiple text boxes.

Among them, the CTPN (Connectionist Text Proposal Network) network model, also known as the Connectionist Text Proposal Network model, is a fully convolutional network that allows input images of any size. It detects text lines by densely sliding small windows in a convolutional feature map, and outputs a series of fine-grained (e.g., fixed width of 16 pixels) text proposals, which are then directly processed by a series of fine-grained text proposals in the convolutional feature map. Detect lines of text in text proposals. The CTPN network model is equipped with a vertical anchor mechanism, which can jointly predict the precise location and text/non-text score of each proposal, thereby achieving accurate text positioning. Adopting CTPN overcomes some major limitations brought by the bottom-up approach based on character detection in related technologies; such as the use of sliding windows in the convolutional layer allows shared convolution computation, which is more expensive than in related technologies. The sliding window-based method greatly reduces the amount of calculation. And through the design of vertical anchor points and fine-grained detection strategies, single-scale images can be used to process text lines of various scales and aspect ratios, thus further reducing the amount of calculation and predicting the accurate position of text lines.

In an optional implementation, when obtaining the position information of the text box in the ID card image, the following steps 1 to 4 may be referred to:

Step 1, obtain the vertex coordinates of the text box in the ID card image. For example, obtain the vertex coordinates of the text box. In one embodiment, the same vertex coordinates of each text box can be uniformly obtained, such as obtaining the vertex coordinates of the upper left corner of the text box. In another embodiment, the four vertex coordinates of the upper left corner, lower left corner, upper right corner, and lower right corner of each text box can be obtained.

Step 2: Determine whether the text box is tilted based on the vertex coordinates of the text box.

In one implementation, whether each text box is tilted can be determined individually. Specifically, if you obtain the four vertex coordinates of the text box, you can determine whether the text box is tilted based on the four vertex coordinates of each text box. For example, if you determine the upper left corner vertex and the lower left corner of the text box, If the abscissas of the corner vertices are the same, and the ordinates of the upper-left vertex are the same as the ordinates of the upper-right vertex, then it is determined that the text box is not tilted. On the contrary, if there is a difference, it means that the text box is tilted.

In another implementation, whether the text box is tilted can be comprehensively determined based on the positions of multiple text boxes. Specifically, if the same vertex coordinates of the text box are obtained, whether the text box is tilted can be determined based on horizontally adjacent text boxes and/or vertically adjacent text boxes. Specifically, for ease of understanding, refer to the schematic diagram of a tilted ID card shown in FIG. 2 , which shows a tilted ID card image 200 and multiple tilted text boxes on the ID card image 200 ( FIG. 2 4 text boxes are simply illustrated in the figure, namely the first text box 202a, the second text box 202b, the third text box 204a, and the fourth text box 204b). Wherein, the shape of the aforementioned text box is a rectangle, and the coordinates of the upper left corner vertex of each text box are selected to represent the position of the text box. In addition, in order to more intuitively display the relationship between the vertex coordinates of each text box, Figure 2 also shows the plane rectangular coordinate axis; where the x-axis is the horizontal axis; the y-axis is the vertical axis; 0 represents the origin. The vertical coordinates of the vertices of the first text box 202a and the second text box 202b differ by 2by-2ay; the abscissa coordinates of the vertices of the first text box 202a and the third text box 204a differ by 4ax-2ax; the third text box 204a and the fourth text box The vertical coordinate difference of the vertex of 204b is 4by-4ay; the horizontal coordinate difference of the vertex of the fourth text box 204b and the second text box 202b is 4bx-2bx.

It can be understood that if the text box is not tilted, the values of 2by-2ay, 4by-4ay, 4ax-2ax, and 4bx-2bx should all be 0. If they are not zero, it proves that the text box is skewed.

It should be noted that Figure 2 is only a brief illustration. It simply shows the name and gender, but does not show the address, ethnicity, date of birth and other information in detail. Figure 2 is a simple schematic diagram of an ID card. It does not correspond to the layout on the actual ID card, nor should it be considered a limitation.

Step 3: If the text box is tilted, adjust the text box.

In an optional implementation, the position of each text box can be adjusted individually, so that the abscissa coordinate of the upper left corner vertex of the adjusted text box is the same as the abscissa coordinate of the lower left corner vertex, and the ordinate coordinate of the upper left corner vertex is the same as the upper right corner vertex. The vertical coordinates of the vertices are the same. In specific implementation, a reference text box (such as the above-mentioned first text box 202a) can be adjusted first, and then the vertical coordinate of the vertex of the text box horizontally adjacent to the reference text box is adjusted to the vertical coordinate of the vertex of the reference text box. The coordinates are the same, and the horizontally adjacent text box is adjusted based on the adjusted vertex of the horizontally adjacent text box; similarly, the abscissa coordinate of the vertex of the text box vertically adjacent to the reference text box is adjusted to The horizontal coordinates of the vertices of the reference text box are the same, and the adjusted vertices of the vertically adjacent text boxes are used as the basis to adjust the vertically adjacent text boxes. Through the above method, you can adjust the position of each text box on the ID card image. By setting the vertex coordinates of a certain text box as a reference point to adjust the vertex coordinates of other text boxes, the efficiency of adjusting text boxes can be effectively improved.

Specifically, for ease of understanding, refer to FIG. 3 , which is a schematic diagram of an adjusted ID card according to an embodiment of the present invention. On the basis of Figure 2, taking the vertex coordinates of the upper left corner of the text box to represent the vertex coordinates of the text box as an example, and using the vertex coordinates of the upper left text box (such as the above-mentioned first text box 202a) as a reference, adjust each text box. Positive, the ordinate of the vertex of the second text box 202b is adjusted to be consistent with the ordinate of the vertex of the first text box 202a, both are 2ay; the abscissa of the vertex of the third text box 204a is adjusted to the abscissa of the vertex of the first text box 202a. are consistent, both are 2ax; correspondingly, the ordinate of the vertex of the fourth text box 204b is adjusted to be consistent with the ordinate of the vertex of the third text box 204a, both are 4ay; the abscissa of the vertex of the fourth text box 204b is consistent with that of the second text box The abscissa coordinates of the vertices of 202b are adjusted to be consistent, both are 2bx. Each text box also adjusts its position based on its adjusted vertex coordinates. The abscissa coordinate of the upper left vertex of the adjusted text box is the same as the abscissa of the lower left vertex, and the ordinate of the upper left vertex is the same as the ordinate of the upper right vertex. The coordinates are the same.

Step 4, obtaining the position information of the adjusted text box.

Similarly, you can obtain the coordinates of one or more vertices of each adjusted text box or the coordinates of key points on the border lines, etc.

After detecting the text box, text recognition can be performed on the text box to obtain text information corresponding to the text box. This embodiment provides a specific text recognition method: text recognition is performed on the text box through the CRNN network model to obtain the text information corresponding to the text box.

Among them, the CRNN (Convolutional Recurrent Neural Network) network model, also known as the convolutional recurrent neural network, is an end-to-end text recognition network model. Can be used with CTPN network models, including convolutional layers, recurrent layers and transcription layers. At the bottom of CRNN, convolutional layers automatically extract feature sequences from each input image. On top of the convolutional network, a recurrent network is constructed to predict each frame of the feature sequence output by the convolutional layer. A transcription layer on top of the CRNN is used to convert the per-frame predictions of the recurrent layer into a label sequence. Although CRNN is composed of different types of network architectures (such as CNN and RNN), it can be jointly trained end-to-end through a loss function. Using the CRNN network model to perform text recognition on text boxes can obtain text recognition results more accurately and reliably.

Furthermore, considering that there may be interfering text other than ID card information on the ID card image to be recognized, such as "This ID card is only used for XX occasions" and other impurity information written in the blank space or watermarks, etc., the text boxes of these interfering texts The position is different from the legal position of the ID card information text box. The embodiment of the present invention can filter the interfering text by judging that the position of the text box of the interfering text is illegal. Specifically, in an optional method, before performing the step of character recognition of the text box, it can be determined whether there is text with illegal position based on the position information of the text box and the legal position of the preset ID card text box. box; if so, delete the text box with illegal position. In another alternative method, before determining the corresponding text information, the illegal text information on the ID card image can be identified and filtered based on the position information and text information of the text box; wherein, the text where the illegal text information is located The position of the box does not belong to the legal position of the preset ID card text box.

In addition, the embodiment of the present invention also provides a second method for identifying ID card information. This method is implemented on the basis of the method shown in Figure 1. This method further describes that based on the position information and text information of the text box, it is determined that the For the specific process of corresponding text information, see Figure 4. The method includes the following steps:

Step S402, obtaining an ID card image of the ID card information to be identified;

Step S404, perform text line detection on the ID card image to obtain multiple text boxes;

Step S406, obtain the position information of the text box in the ID card image;

Step S408: Perform text recognition on the text box to obtain text information corresponding to the text box;

Step S410, determine the text box of the same peer according to the position information of the text box;

The characteristic information of the distribution position of the text boxes in the same row on the ID card can be used in actual application, such as the characteristic information of the ordinates of the corresponding coordinate points on the text boxes in the same row in the plane rectangular coordinate system where multiple text boxes coexist. For example, regardless of whether the ID card image is tilted or not, the spacing between the ordinates of the corresponding coordinate points of the text boxes in the same row is within the preset threshold range. For example, if the ID card image is not tilted, the ordinates of the upper left corner vertices of the two text boxes in the same row should be the same. If the ID card image is tilted, the ordinates of the upper left corner vertices of the two text boxes in the same row will deviate, but they will also be within a certain range. Alternatively, the ordinate value range of the vertical borders of the text boxes in the same row will have a certain degree of overlap.

Step S412: Determine text boxes with corresponding relationships among the text boxes of the same party based on the position information of the text boxes of the same party;

For example, the abscissa coordinates of the text boxes in the same row in the plane rectangular coordinate system can be analyzed, and the text boxes with a small difference in abscissa coordinates, that is, close to each other, can be determined as text boxes with a corresponding relationship.

Step S414: Determine the text information that has the corresponding relationship based on the text box that has the corresponding relationship and the text information of the text box.

Whether the corresponding relationship of the text box is correct can be checked based on the text box with the corresponding relationship and the text information of the text box; the text information of the text box with the correct corresponding relationship can be determined as the text information with the corresponding relationship. For example, the text boxes where "gender" and "female" are located are similar, and the text boxes where "ethnicity" and "Han" are located are similar. In step S412, it is determined that the text boxes have a corresponding relationship, and it is determined that "gender" and "female" are verified through text information verification. ” is text information with a corresponding relationship, and “ethnicity” and “Han” are text information with a corresponding relationship. Since "female" and "ethnicity" are also adjacent in the ID card image, if step S412 determines that the text box where "female" and "ethnicity" are located has a corresponding relationship, then in step S414, it can be determined through text information verification that " There is no corresponding relationship between "female" and "ethnic".

Step S416: Determine the ID card information corresponding to the ID card image based on the text information with the corresponding relationship.

The above-mentioned second identification method of ID card information provided by the embodiment of the present invention first determines the text box of the peer based on the position information of the text box, and then determines that the text box of the peer has a corresponding relationship based on the position information of the text box of the peer. the text box; and then determine the text information with the corresponding relationship based on the text box with the corresponding relationship and the text information of the text box. The above method provided by the embodiment of the present invention helps to improve the accuracy of identification of ID card information by first determining the corresponding relationship of the text box, and then combining the text information of the text box to determine the text information with the corresponding relationship.

When obtaining the position information of the text box in the ID card image and determining the text boxes in the same line and the text boxes in the same line with corresponding relationships, the following two main implementation methods can be referred to:

method one:

The obtained position information of the above text box includes the starting point coordinates of the text box, and the starting point coordinates include the starting abscissa and the starting ordinate; among them, most text boxes are rectangles, and the starting point can be the coordinates of any vertex of the text box. , for example, the coordinates of the upper left corner vertex of the text box can be used as the starting point coordinates.

The above step S410 may be to determine the text boxes in the same line by the starting ordinate of the text box, specifically, including the following steps: determining the text boxes whose starting point ordinates differ by a preset first threshold as the text boxes in the same line.

Considering that the actually obtained ID card image is skewed, the multiple text boxes obtained by text line detection on the ID card image may be skewed, causing the vertical coordinate values of the text boxes in the same row to be skewed on the plane rectangular coordinate system. Therefore, the first threshold can be set in advance. When the difference value of the ordinates of the starting points of multiple text boxes is less than the preset first threshold, the multiple text boxes are determined to be text boxes of the same row. The preset first threshold can be set based on empirical values and is not limited here.

Based on this, the above-mentioned step S412, that is, the step of determining text boxes with corresponding relationships in the text boxes of the same party based on the position information of the text boxes of the same party, includes: setting the abscissa difference of the starting points in the text boxes of the same party by a preset second second. The text boxes within the threshold are determined to be text boxes with corresponding relationships among the text boxes of the same row.

Among them, the distance between corresponding text boxes in the same peer group on the actual ID card is smaller than the distance between text boxes that have no corresponding relationship. That is, the corresponding text boxes in the same peer group are relatively similar in location distribution. Therefore, the difference between the abscissa coordinates of the starting points in the text boxes of the same row is used as the distance between the text boxes, and compared with the preset second threshold, the text boxes within the preset second threshold of the difference between the abscissa coordinates of the starting points in the text boxes of the same row are determined. It is a text box with a corresponding relationship among the text boxes of the same row. The preset second threshold can be set based on empirical values and is not limited here.

Method two:

The obtained position information of the above-mentioned text box includes the ordinate value range and the abscissa value range of the text box; among them, most text boxes are rectangles, and the ordinate value range and abscissa value range of the text box can be determined according to the text box. Determined by the coordinates of the four vertices.

The above step S410 may be to determine the text boxes of the same row according to the ordinate value range of the text box. Specifically, it includes the following steps: determining two text boxes whose ordinate value range has at least a first overlapping length as the text boxes of the same row. . For example, if the ID card is not skewed, the ordinate value range of the first text box (that is, the vertical edge of the first text box is mapped to the value range on the Y-axis) is the ordinate value range of the second text box in the same row. The value range should be the same (that is, completely overlapping). If the ID card is skewed, the text boxes in the same row will also be skewed accordingly (see Figure 2 for details), but the ordinate value range of the first text box is the same as that of the second text box. There will still be a certain degree of overlap in the ordinate value range of the text boxes.

Considering that the actually obtained ID card image is skewed, the multiple text boxes obtained by text line detection on the ID card image may be skewed, causing the ordinate value ranges of the same text boxes on the plane rectangular coordinate system to not completely overlap. , but there is a certain deviation, so the first overlapping length is preset. When the overlapping area of the ordinate value range of the two text boxes is greater than the first overlapping length, the two text boxes are determined to be in the same row. The first overlap length can be set based on empirical values and is not limited here.

Based on this, the above-mentioned step S412, that is, the step of determining text boxes with corresponding relationships among the text boxes of the same party based on the position information of the text boxes of the same party, includes: determining the text box according to the abscissa value range of the text boxes of the same party. The minimum value of the abscissa coordinate of the text box; specifically, the minimum value of the abscissa coordinate of the text box is the abscissa coordinate of the upper left corner vertex or the lower left corner vertex abscissa of the text box. The text boxes in the text boxes of the same row that differ within a preset third threshold value from the minimum value of the abscissa in the text boxes of the same row are determined to be text boxes with a corresponding relationship in the text boxes of the same row.

Among them, the distance between corresponding text boxes in the same peer group on the actual ID card is smaller than the distance between text boxes that have no corresponding relationship. That is to say, the position distribution of the corresponding text boxes in the same peer group is relatively similar. Therefore, the difference between the minimum abscissa coordinates of the text boxes in the same row is used as the distance between the text boxes, and compared with the preset third threshold, the text boxes within the preset third threshold of the starting point abscissa difference of the text boxes in the same row are determined. It is a text box with a corresponding relationship among the text boxes of the same row. The preset third threshold can be set based on experience values and is not limited here.

The embodiment of the present invention also provides a third method for identifying ID card information, which is implemented on the basis of the method shown in FIG. 1 . The method further describes a specific process of determining the ID card information corresponding to the ID card image according to the text information having a corresponding relationship. Referring to FIG. 5 , the method includes the following steps:

Step S502, obtain the ID card image of the ID card information to be recognized;

Step S504: Perform text line detection on the ID card image to obtain multiple text boxes;

Step S506, obtaining the position information of the text box in the ID card image;

Step S510, determine text information with a corresponding relationship based on the position information and text information of the text box;

Step S512: Based on the text information with the corresponding relationship, determine the information category of the text information with the corresponding relationship;

For example, the information category includes one or more of name category, gender category, ethnicity category, date of birth category, address category and ID number category.

Step S514, determining the key words in the text information having a corresponding relationship and the text having a corresponding relationship with the key words according to the information category;

Among them, the key words include one or more of name, gender, ethnicity, birth, year, month, day, address and citizen ID number.

Step S516, determining the text corresponding to the keyword as the ID card information.

Specifically, for example, the keyword "name" corresponds to "Zhang San"; the keyword "gender" corresponds to "male"; the keyword "ethnicity" corresponds to "Han", then the text "Zhang San-male-Han" that has a corresponding relationship with the above keywords is extracted and determined as the ID card information.

The above-mentioned third identification method of ID card information provided by the embodiment of the present invention first determines the information category of the text information with the corresponding relationship, and then determines the key words in the text information with the corresponding relationship and the keywords according to the information category. The text that has a corresponding relationship with the keyword; then the text that has a corresponding relationship with the key word is determined as the ID card information. The above method provided by the embodiment of the present invention can extract valid ID card information by determining the key words in the text information with corresponding relationships according to the information category, and determining the text with the corresponding relationship with the key words as ID card information. Helps improve the accuracy of identification of ID card information.

Further considering the practical application, there is an ID card image to be recognized that may be an illegal ID card image. The above-mentioned third identification method of ID card information provided by the embodiment of the present invention also includes: checking the ID card according to the information category. Check whether the information is correct; when the result of the check is no, an error message will be issued. There are many ways to report errors. For example, you can directly leave the wrong ID card information blank, or fill in the specified value "Error" or other values that represent the error, or you can also directly highlight the wrong ID card information. For marking, any form of error prompting method is acceptable, and there are no restrictions here.

In specific implementation, according to the information category, you can check whether the ID card information is correct by executing the following judgment conditions:

Judgment condition 1: Use the information corresponding to the ID card number category in the ID card information to check whether the information corresponding to the date of birth category in the ID card information and the information corresponding to the gender category in the ID card information are correct. :For example, the 11-12 digits and 13-14 digits of the ID card number represent the year, month and day of birth respectively. You can check whether the date of birth in the ID card information is correct by using the specified digits on the ID card number. In addition, the ID card number The odd or even value of a specified position on the number represents gender, the odd number represents male, and the even number represents female. Through the odd or even value of the specified position, you can check whether the gender in the ID card information is correct.

Judgment condition 2: Check whether the information corresponding to the gender category in the ID card information is one of the preset gender words, where the preset gender words include male or female;

Judgment condition 3: Check whether the information corresponding to the ethnic category in the ID card information is one of the preset ethnic words, where the preset ethnic words include legal ethnic words, such as Han, Hui, Manchu and Uyghur;

Judgment condition 4: Check whether the information corresponding to the date of birth category in the ID card information is one of the preset numbers; where the preset number is a legal date number. For example, the legal date numbers for months are 1 to 12, the legal date numbers for individual days are 1 to 31, etc.

During specific implementation, one or more of the above judgment conditions can be selected to check whether the ID card information is correct or not.

In addition, the identification method of ID card information provided in this embodiment also includes: determining whether the ID card image is legal based on the information corresponding to the ID card number category in the ID card information.

For example, it can be checked whether the ID number is legal according to the numbering rules of the ID number. If not, the ID card image corresponding to the ID number category is illegal.

Specifically, the ID card number has 18 digits, and the numbering rules from left to right are as follows: digits 1-2 represent provincial administrative district codes; digits 3-4 represent prefecture-level administrative district codes; digits 5-6 represent county administrative district codes. ; Digits 7-10, 11-12 and 13-14 respectively represent the year, month and day of birth; digits 15-17 represent the sequence code, which is the number of a person born in the same year, month and day in the same area, among which the odd number is a male , even numbers are women; 18 digits are the check code, by multiplying the first 17 digits of the ID number by different coefficients; among them, the coefficients from the 1st to the 17th digit are: 7-9-10-5 -8-4-2-1-6-3-7-9-10-5-8-4-2; add the result of multiplying the 17-digit number and the coefficient, and then divide it by 11, the remainder is is the aforementioned check code, in which, when the remainder is 0-9, it is represented by 0-9, and when the remainder is 10, it is represented by X (Roman numeral 10). During specific implementation, the above encoding rules can be used to verify whether the extracted ID number is legal.

Corresponding to the first identification card information recognition method described above, an embodiment of the present invention provides a first identification card information recognition device, including:

The image to be recognized acquisition module 602 is used to obtain the ID card image of the ID card information to be recognized;

The text line detection module 604 is used to detect text lines on the ID card image to obtain multiple text boxes;

The location information acquisition module 606 is used to obtain the location information of the text box in the ID card image;

The text recognition module 608 is used to perform text recognition on the text box and obtain the text information corresponding to the text box;

Correspondence determination module 610 is used to determine text information with corresponding relationships based on the position information and text information of the text box;

The ID card information determination module 612 is used to determine the ID card information corresponding to the ID card image according to the text information having a corresponding relationship.

The present invention provides a first identification device for ID card information, which can first perform text line detection on an ID card image to be recognized to obtain multiple text boxes; and then obtain the position information of the text box and the text box in the ID card image. The corresponding text information is determined based on the position information and text information of the text box, and the corresponding text information is determined, and the ID card information corresponding to the ID card image is determined based on the corresponding text information. Unlike the existing technology, which mostly directly recognizes the text on the ID card image to extract the ID card information, the wrong ID card information may be extracted due to reasons such as scattered text on the ID card image, watermarks added or the ID card image being tilted. Compared with the problem, the embodiment of the present invention can first detect the text box on the ID card image, and then determine the corresponding text information based on the position information and text information of the text box, which can avoid extracting the wrong ID card information and effectively Improved the recognition accuracy of ID card information.

In one implementation, the text line detection module includes the following units:

The ID card tilt determination unit is used to obtain the vertex coordinates of the ID card image and determine whether the ID card image is tilted based on the vertex coordinates of the ID card image;

An image correction unit, used for correcting the ID card image when the judgment result of the ID card tilt judgment unit is yes;

The horizontal coordinate of the upper left corner vertex of the adjusted ID card image is the same as the horizontal coordinate of the lower left corner vertex, the vertical coordinate of the upper left corner vertex is the same as the vertical coordinate of the upper right corner vertex, and the distance between the upper left corner vertex and the upper right corner vertex is greater than the distance between the upper left corner vertex and the lower left corner vertex;

The text line detection unit is used to detect text lines on the corrected ID card image to obtain multiple text boxes.

In a specific implementation, the above text line detection module is also used to detect text lines on the ID card image through the CTPN network model to obtain multiple text boxes.

In one implementation, the above-mentioned position information acquisition module includes: a coordinate acquisition unit, which acquires the vertex coordinates of the text box in the ID card image; a text frame tilt determination unit, used to determine whether the text box is tilted based on the vertex coordinates of the text box; The text box adjustment unit is used to adjust the text box when the judgment result of the text box tilt judgment unit is yes; the position information acquisition unit is used to obtain the position information of the adjusted text box.

In a specific implementation, the above-mentioned text information determination module is also used to perform text recognition on the text box through a CRNN network model to obtain text information corresponding to the text box.

Furthermore, the above-mentioned first ID card information identification method device also includes an illegal text box deletion module, which is used to, before performing text recognition on the text box, based on the position information of the text box and the preset legal position of the ID card text box, Determine whether there is a text box with an illegal position; when there is a text box with an illegal position, delete the text box with an illegal position.

In addition, the above-mentioned first ID card information identification method device also includes an illegal text information filtering module, which is used to identify and filter the illegal text information on the ID card image based on the position information and text information of the text box before determining the text information with the corresponding relationship. The illegal text information; among them, the position of the text box where the illegal text information is located does not belong to the legal position of the preset ID card text box.

Corresponding to the above-mentioned second identification method of ID card information, embodiments of the present invention provide a second identification device of ID card information. Based on the device shown in Figure 6, this device further shows the corresponding relationship determination module 610. For the specific structure, see Figure 7. The correspondence determination module 610 includes:

The peer text box determination unit 702 is used to determine the peer text box according to the position information of the text box;

The corresponding text box determining unit 704 is configured to determine, based on the position information of the text boxes of the same party, the text boxes that have a corresponding relationship among the text boxes of the same party.

The corresponding text information determining unit 706 is used to determine text information with a corresponding relationship based on the text boxes with the corresponding relationship and the text information of the text boxes.

The above-mentioned second identification device for ID card information provided by the embodiment of the present invention first determines the text box of the peer based on the position information of the text box, and then determines that the text box of the peer has a corresponding relationship based on the position information of the text box of the peer. the text box; and then determine the text information with the corresponding relationship based on the text box with the corresponding relationship and the text information of the text box. The embodiment of the present invention helps to improve the accuracy of identification of ID card information by first determining the corresponding relationship of the text box, and then combining the text information of the text box to determine the text information with the corresponding relationship.

In a specific implementation, when the position information of the text box includes the starting point coordinates of the text box, and the starting point coordinates include the starting abscissa and the starting ordinate, the peer text box determination unit is also used to change the starting point vertically. Text boxes whose coordinates differ within a preset first threshold are determined to be text boxes of the same row. Based on this, the above-mentioned corresponding text box determination unit is also used to determine the text boxes in the text boxes of the same row whose starting point abscissas differ within a preset second threshold as the text boxes in the text boxes of the same row that have a corresponding relationship.

In another specific embodiment, when the position information of the above-mentioned text box includes the vertical coordinate value range and the horizontal coordinate value range of the text box, the same-line text box determination unit is also used to determine two text boxes whose vertical coordinate value ranges have at least a first overlapping length as text boxes in the same line; based on this, the above-mentioned corresponding text box determination unit is also used to determine the minimum horizontal coordinate value of the text box according to the horizontal coordinate value range of the text boxes in the same line; and determine the text boxes whose horizontal coordinate minimum values among the text boxes in the same line differ by a preset third threshold as text boxes in the same line with corresponding relationships.

The above-mentioned corresponding text information determination unit is also used to check whether the corresponding relationship of the text box is correct based on the text box with the corresponding relationship and the text information of the text box; and determine the text information of the text box with the correct corresponding relationship as having the corresponding relationship. text information.

Corresponding to the above-mentioned third identification method of ID card information, the implementation force of the present invention provides a third identification device of ID card information. Based on the device shown in Figure 6, this device further shows an ID card information determination module 612. For the specific structure, see Figure 8. The ID card information determination module 612 includes:

The information category determination unit 802 is used to determine the information category of the text information with the corresponding relationship based on the text information with the corresponding relationship; the information category includes the name category, gender category, ethnic category, date of birth category, address category and ID number category. one or more of;

The key words and corresponding text determination unit 804 is used to determine the key words in the text information with corresponding relationships and the text with corresponding relationships with the key words according to the information category; wherein the key words include name, gender, ethnicity , one or more of birth, year, month, day, address and citizen ID number;

The identity card information determining unit 806 is used to determine the text corresponding to the keyword as the identity card information.

The above-mentioned third identification device for ID card information provided by the embodiment of the present invention first determines the information category of the text information with the corresponding relationship, and then determines the keywords in the text information with the corresponding relationship and the keywords based on the information category. The text that has a corresponding relationship with the keyword; then the text that has a corresponding relationship with the key word is determined as the ID card information. The embodiment of the present invention determines the key words in the text information with corresponding relationships according to the information category, and determines the words corresponding to the key words as ID card information, which can extract valid ID card information and help improve the quality of the information. The accuracy of identification of ID card information.

Furthermore, the above-mentioned third identification device for ID card information also includes a checking module and an error reporting module; wherein, the checking module is used to check whether the ID card information is correct according to the information category; and the error reporting module is used when the judgment result of the checking module is If not, an error message will be issued.

Specifically, the above-mentioned inspection module is also used to use the information corresponding to the ID card number category in the ID card information to check whether the information corresponding to the date of birth category in the ID card information and the information corresponding to the gender category in the ID card information are correct; and/or, to check whether the information corresponding to the gender category in the ID card information is one of the preset gender words, wherein the preset gender words include male or female; and/or, to check whether the information corresponding to the ethnic category in the ID card information is one of the preset ethnic words, wherein the preset ethnic words include legal ethnic words; and/or, to check whether the information corresponding to the date of birth category in the ID card information is one of the preset numbers; wherein the preset numbers are legal date numbers.

Furthermore, the above-mentioned third identification device for ID card information also includes a legality judgment module, which is used to determine whether the ID card image is legal based on the information corresponding to the ID card number category in the ID card information.

The implementation principle and technical effects of the above-mentioned device provided in this embodiment are the same as those in the previous embodiment. This is a brief description. For matters not mentioned in the device embodiment, please refer to the corresponding content in the above-mentioned method embodiment.

Furthermore, this embodiment also provides a terminal device, which includes a processor and a memory, wherein the memory is used to store one or more computer instructions, and the one or more computer instructions are executed by the processor to realize the above identity. How to identify certification information.

Referring to the schematic structural diagram of a terminal device shown in Figure 9, the terminal device 900 is shown, including: a processor 90, a memory 91, a bus 92 and a communication interface 93. The processor 90, the communication interface 93 and the memory 91 pass through the bus 92 Connection; processor 90 is used to execute executable modules stored in memory 91, such as computer programs.

The memory 91 may include high-speed random access memory (RAM), or may also include non-volatile memory (non-volatile memory), such as at least one disk memory. The communication connection between the system network element and at least one other network element is realized through at least one communication interface 93 (which can be wired or wireless), and the Internet, wide area network, local network, metropolitan area network, etc. can be used.

The bus 92 may be an ISA bus, a PCI bus, an EISA bus, etc. The bus can be divided into address bus, data bus, control bus, etc. For ease of presentation, only one bidirectional arrow is used in Figure 9, but it does not mean that there is only one bus or one type of bus.

The memory 91 is used to store the program. The processor 90 executes the program after receiving the execution instruction. The method executed by the device for stream process definition disclosed in any of the embodiments of the present invention can be applied to processing. in the processor 90, or implemented by the processor 90.

The processor 90 may be an integrated circuit chip with signal processing capabilities. During the implementation process, each step of the above method can be completed by instructions in the form of hardware integrated logic circuits or software in the processor 90 . The above-mentioned processor 90 can be a general-purpose processor, including a Central Processing Unit (CPU for short), a Network Processor (NP for short), etc.; it can also be a Digital Signal Processing (DSP for short) , Application Specific Integrated Circuit (ASIC for short), Field-Programmable Gate Array (FPGA for short) or other programmable logic devices, discrete gate or transistor logic devices, and discrete hardware components. Each method, step and logical block diagram disclosed in the embodiment of the present invention can be implemented or executed. A general-purpose processor may be a microprocessor or the processor may be any conventional processor, etc. The steps of the method disclosed in conjunction with the embodiments of the present invention can be directly implemented by a hardware decoding processor, or executed by a combination of hardware and software modules in the decoding processor. The software module can be located in random access memory, flash memory, read-only memory, programmable read-only memory or electrically erasable programmable memory, registers and other mature storage media in this field. The storage medium is located in the memory 91. The processor 90 reads the information in the memory 91 and completes the steps of the above method in combination with its hardware.

Furthermore, embodiments of the present invention also provide a computer-readable storage medium. A computer program is stored on the computer-readable storage medium. When the computer program is run by a processor, the steps of any of the above identification methods for ID card information are executed. For specific implementation, please refer to the method embodiments and will not be described again here.

The identification method, device, and computer program product for terminal equipment provided by the embodiments of the present invention include a computer-readable storage medium storing program code. The instructions included in the program code can be used to execute the previous method embodiments. The specific implementation of the method can be found in the method embodiments and will not be described again here.

In addition, in the description of the embodiments of the present invention, unless otherwise clearly stated and limited, the terms "installation", "connection" and "connection" should be understood in a broad sense. For example, it can be a fixed connection or a detachable connection. , or integrally connected; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium; it can be an internal connection between two components. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood on a case-by-case basis.

If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention essentially or the part that contributes to the existing technology or the part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including Several instructions are used to cause a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage media include: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), magnetic disk or optical disk and other media that can store program code. .

In the description of the present invention, it should be noted that the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc., indicating the orientation or positional relationship, are based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, the terms "first", "second", and "third" are used for descriptive purposes only, and cannot be understood as indicating or implying relative importance.

Finally, it should be noted that the above-mentioned embodiments are only specific implementations of the present invention and are used to illustrate the technical solutions of the present invention rather than to limit them. The protection scope of the present invention is not limited thereto. Although refer to the foregoing The embodiments illustrate the present invention in detail. Those of ordinary skill in the art should understand that any person familiar with the technical field can still modify the technical solutions recorded in the foregoing embodiments within the technical scope disclosed by the present invention. It may be easy to think of changes, or equivalent substitutions of some of the technical features; and these modifications, changes or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and they should all be included in the present invention. within the scope of protection. Therefore, the protection scope of the present invention should be subject to the protection scope of the claims.

Claims (17)

1. The identification method of the identity card information is characterized by comprising the following steps:

acquiring an identity card image of identity card information to be identified;

performing text line detection on the identity card image to obtain a plurality of text boxes;

acquiring the position information of the text box in the identity card image;

Performing character recognition on the text box to obtain character information corresponding to the text box;

determining text information with a corresponding relation according to the position information and the text information of the text box;

according to the text information with the corresponding relation, determining the identity card information corresponding to the identity card image;

the step of determining the text information with the corresponding relation according to the position information and the text information of the text box comprises the following steps:

determining the text boxes of the same line according to the position information of the text boxes;

determining text boxes with corresponding relations in the text boxes of the same line according to the position information of the text boxes of the same line;

and determining the text information with the corresponding relation according to the text box with the corresponding relation and the text information of the text box.

2. The method of claim 1, wherein the step of performing text line detection on the identification card image to obtain a plurality of text boxes comprises:

acquiring vertex coordinates of the identity card image, and judging whether the identity card image is inclined or not according to the vertex coordinates of the identity card image;

if yes, the identity card image is regulated;

And detecting the text lines of the aligned identity card images to obtain a plurality of text boxes.

3. The method of claim 1, wherein the step of performing text line detection on the identification card image to obtain a plurality of text boxes comprises:

and detecting the text lines of the identity card image through a CTPN network model to obtain a plurality of text boxes.

4. The method of claim 1, wherein the step of obtaining location information of the text box in the identification card image comprises:

obtaining vertex coordinates of the text box in the identity card image;

judging whether the text box is inclined or not according to the vertex coordinates of the text box;

if yes, the text box is adjusted;

and acquiring the position information of the aligned text box.

5. The method of claim 1, wherein the step of performing text recognition on the text box to obtain text information corresponding to the text box comprises:

and carrying out character recognition on the text box through the CRNN network model to obtain character information corresponding to the text box.

6. The method of claim 1, wherein the location information of the text box comprises a start point coordinate of the text box, the start point coordinate comprising a start abscissa and a start ordinate;

The step of determining the text boxes of the same line according to the position information of the text boxes comprises the following steps:

determining the text boxes within a preset first threshold value of the difference of the ordinate of the starting point as the text boxes of the same row;

the step of determining the text boxes with the corresponding relation in the text boxes of the same line according to the position information of the text boxes of the same line comprises the following steps:

and determining the text boxes in the starting point abscissa differences among the text boxes in the same row within a preset second threshold as the text boxes with corresponding relations among the text boxes in the same row.

7. The method of claim 1, wherein the location information of the text box includes an ordinate range and an abscissa range of values of the text box;

the step of determining the text boxes of the same line according to the position information of the text boxes comprises the following steps:

determining two text boxes with at least a first overlapping length in the ordinate value range as the text boxes in the same row;

the step of determining the text boxes with the corresponding relation in the text boxes of the same line according to the position information of the text boxes of the same line comprises the following steps:

determining the minimum value of the abscissa of the text boxes according to the value range of the abscissa of the text boxes in the same row;

And determining the text boxes in the minimum difference of the abscissa values in the text boxes in the same row within a preset third threshold as the text boxes with corresponding relations in the text boxes in the same row.

8. The method according to claim 1, wherein the step of determining text information having a correspondence relationship from the text box having a correspondence relationship and text information of the text box includes:

checking whether the corresponding relation of the text box is correct or not according to the text box with the corresponding relation and the text information of the text box;

and determining the text information of the text box with the correct corresponding relation as the text information with the corresponding relation.

9. The method of claim 1, wherein the step of determining the identification card information corresponding to the identification card image according to the text information having the correspondence relationship comprises:

determining information category of the text information with the corresponding relation based on the text information with the corresponding relation; the information category comprises one or more of a name category, a gender category, a ethnicity category, a birth date category, an address category and an identification card number category;

According to the information category, determining key words in the text information with the corresponding relation and text with the corresponding relation with the key words; wherein the key words comprise one or more of name, gender, ethnicity, birth, year, month, day, address and citizen ID card number;

and determining the characters with the corresponding relation with the key words as identity card information.

10. The method according to claim 9, wherein the method further comprises:

checking whether the identity card information is correct according to the information category;

if not, initiating error reporting prompt.

11. The method of claim 10, wherein the step of checking whether the identification card information is correct based on the information category comprises:

checking whether information corresponding to the birth date category in the identity card information and information corresponding to the neutral category in the identity card information are correct or not by adopting the information corresponding to the identity card number category in the identity card information;

and/or the number of the groups of groups,

checking whether the information corresponding to the gender category in the identity card information is one of preset gender words, wherein the preset gender words comprise men or women;

And/or the number of the groups of groups,

checking whether the information corresponding to the national category in the identity card information is one of preset national words, wherein the preset national words comprise legal national words;

and/or the number of the groups of groups,

checking whether the information corresponding to the birth date category in the identity card information is one of preset numbers or not; wherein the preset number is a legal date number.

12. The method according to claim 10, wherein the method further comprises:

judging whether the identity card image is legal or not according to the information corresponding to the identity card number category in the identity card information.

13. The method of claim 1, wherein prior to the step of text recognition of the text box, the method further comprises:

judging whether a text box with illegal positions exists or not according to the position information of the text box and the legal positions of the preset identity card text boxes;

if yes, deleting the text box with illegal position.

14. The method of claim 1, wherein prior to the step of determining text information having a correspondence, the method further comprises:

identifying and filtering illegal character information on the identity card image according to the position information and the character information of the text box; the position of the text box where the illegal word information is located does not belong to the legal position of the preset identity card text box.

15. An identification device for identification card information, comprising:

the image acquisition module to be identified is used for acquiring an identity card image of the identity card information to be identified;

the text line detection module is used for detecting the text lines of the identity card image to obtain a plurality of text boxes;

the position information acquisition module is used for acquiring the position information of the text box in the identity card image;

the text recognition module is used for recognizing the text box to obtain text information corresponding to the text box;

the corresponding relation determining module is used for determining text information with a corresponding relation according to the position information and the text information of the text box;

the identity card information determining module is used for determining identity card information corresponding to the identity card image according to the text information with the corresponding relation;

the correspondence determining module includes:

the same-line text box determining unit is used for determining the text boxes of the same line according to the position information of the text boxes;

the corresponding text box determining unit is used for determining text boxes with corresponding relations in the text boxes of the same line according to the position information of the text boxes of the same line;

And the corresponding text information determining unit is used for determining text information with a corresponding relation according to the text box with the corresponding relation and the text information of the text box.

16. A terminal device, characterized in that it comprises a memory for storing a program supporting the processor to perform the method of any one of claims 1 to 14, and a processor configured to execute the program stored in the memory.

17. A computer storage medium storing computer software instructions for use with the method of any one of claims 1 to 14.