Character detection method based on optical character recognition, medium and system

Abstract

The invention discloses a character detection method, medium and system based on optical character recognition, and the method comprises the steps: obtaining a screen picture, and carrying out the binarization processing of the screen picture, so as to generate a binary picture corresponding to the screen picture; performing single character recognition on the binarized picture to obtain frame coordinate information and confidence coefficient corresponding to each character in the binarized picture; filtering each character based on a preset confidence threshold; for each filtered character, judging whether each character is an abnormal character or not based on the frame coordinate information; if yes, marking the abnormal characters; the abnormal characters in the text can be effectively recognized, and the situation that the abnormal characters are deleted by mistake and cannot be read is prevented.

Description

Technical Field

[0001] The present application relates to the field of text detection technology, and in particular to a text detection method, medium and system based on optical character recognition. Background Art

[0002] After decades of development, optical character recognition (OCR) technology has been applied in various fields such as industry, national defense, and life, and has achieved great success in solving the problem of text recognition from images. However, with the optimization of models and algorithms with continuously improved recognition accuracy, OCR still has certain limitations in some scenarios. For example, the text content cannot be read normally because the text is too small or too dense, and the text color is similar to the background color or the color is too light, which makes it difficult to read.

[0003] In the related art, when optical character recognition technology is used to recognize text, the text recognition and judgment is often performed on a whole paragraph of text. In this way, some valid information such as fonts that are too small, too dense or too light in color are often filtered out and cannot be read. Summary of the invention

[0004] The present invention aims to solve one of the technical problems in the related art at least to a certain extent. To this end, one object of the present invention is to provide a text detection method based on optical character recognition, which can effectively identify abnormal text in a text and prevent the abnormal text from being mistakenly deleted and becoming unreadable.

[0005] In a first aspect, the present invention proposes a text detection method based on optical character recognition, comprising: obtaining a screen image, and binarizing the screen image to generate a binarized image corresponding to the screen image; performing single text recognition on the binarized image to obtain bounding box coordinate information and confidence corresponding to each text in the binarized image; filtering each text based on a preset confidence threshold; for each filtered text, determining whether each text is an abnormal text based on the bounding box coordinate information; if so, marking the abnormal text.

[0006] According to the text detection method based on optical character recognition in an embodiment of the present invention, first, a screen image is obtained, and the screen image is binarized to generate a binarized image corresponding to the screen image; then, single text recognition is performed on the binarized image to obtain the frame coordinate information and confidence corresponding to each text in the binarized image; then, each text is filtered based on a preset confidence threshold; then, for each filtered text, whether each text is an abnormal text is determined based on the frame coordinate information; then, if so, the abnormal text is marked; thereby achieving effective recognition of abnormal text in the text and preventing abnormal text from being accidentally deleted and becoming unreadable.

[0007] In some embodiments, each text is filtered based on a preset confidence threshold, including: determining whether the confidence corresponding to the current text is less than the preset confidence threshold; if the confidence corresponding to the current text is less than the preset confidence threshold, filtering the current text.

[0008] In some embodiments, judging whether each text is an abnormal text based on the border coordinate information includes: calculating the actual width and actual height corresponding to the current text according to the border coordinate information; judging whether the actual width is less than a preset width threshold or whether the actual height is less than a preset height threshold; if the actual width is less than the preset width threshold or the actual height is less than the preset height threshold, determining that the current text is an abnormal text; if the actual width is greater than or equal to the preset width threshold and the actual height is greater than or equal to the preset height threshold, calculating the number of binarized pixels within the area range of the current text according to the border information; judging whether the number of binarized pixels is greater than the preset pixel number threshold; if not, determining that the current text is an abnormal text.

[0009] In some embodiments, judging whether each text is an abnormal text based on the border coordinate information includes: obtaining a preset underline thickness value; determining an underline area corresponding to the current text according to the border coordinate information and the preset underline thickness value; detecting the underline area to determine whether there is an underline in the underline area; if so, determining that the current text is an abnormal text.

[0010] In some embodiments, the underline area is detected to determine whether there is an underline in the underline area, including: determining whether there are continuous black pixels in the underline area; if so, calculating the length value of the continuous black pixels; determining whether the length value of the continuous black pixels is greater than a preset length threshold; if so, determining that there is an underline in the underline area.

[0011] In some embodiments, for each filtered text, whether each text is an abnormal text is determined based on the border coordinate information, including: calculating the center point coordinates of the current text according to the border coordinate information, and calculating the actual width value and the actual height value of the current text according to the border coordinate information; calculating the italic angle value corresponding to the current text according to the center point coordinates, the actual width value and the actual height value; determining whether the italic angle value is greater than a preset angle value threshold; if so, determining that the current text is an abnormal text.

[0012] In some embodiments, for each text after filtering, whether each text is an abnormal text is determined based on the border coordinate information, including: determining the text area corresponding to the current text according to the border coordinate information; filtering the text area to remove white pixels in the text area to obtain a preselected pixel set; calculating the binarized average color value according to the preselected pixel set and the binarized image; calculating the average color value of the original image according to the preselected pixel set and the screen image; calculating the difference between the binarized average color value and the average color value of the original image; determining whether the difference is greater than a preset difference threshold; if so, determining that the current text is an abnormal text.

[0013] In a second aspect, an embodiment of the present invention proposes a computer-readable storage medium on which an optical character recognition-based text detection program is stored. When the optical character recognition-based text detection program is executed by a processor, the optical character recognition-based text detection method as described above is implemented.

[0014] In a third aspect, an embodiment of the present invention proposes a text detection system based on optical character recognition, comprising: an acquisition module, the acquisition module is used to acquire a screen image, and binarize the screen image to generate a binary image corresponding to the screen image; a recognition module, the recognition module is used to perform single text recognition on the binary image to obtain the border coordinate information and confidence corresponding to each text in the binary image; a filtering module, the filtering module is used to filter each text based on a preset confidence threshold; a judgment module, the judgment module is used to judge whether each text after filtering is an abnormal text based on the border coordinate information; a marking module, the marking module is used to mark the abnormal text when the text is judged to be an abnormal text.

[0015] According to the text detection system based on optical character recognition of the embodiment of the present invention, an acquisition module is set to acquire a screen image, and the screen image is binarized to generate a binarized image corresponding to the screen image; a recognition module is used to perform single text recognition on the binarized image to obtain the frame coordinate information and confidence corresponding to each text in the binarized image; a filtering module is used to filter each text based on a preset confidence threshold; a judgment module is used to judge whether each text after filtering is an abnormal text based on the frame coordinate information; a marking module is used to mark the abnormal text when the text is judged to be an abnormal text; thereby achieving effective recognition of abnormal text in the text and preventing abnormal text from being accidentally deleted and becoming unreadable.

[0016] In some embodiments, each text is filtered based on a preset confidence threshold, including: determining whether the confidence corresponding to the current text is less than the preset confidence threshold; if the confidence corresponding to the current text is less than the preset confidence threshold, filtering the current text.

[0017] Additional aspects and advantages of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 is a flow chart of a text detection method based on optical character recognition according to an embodiment of the present invention; Figure 2 is a block diagram of a text detection system based on optical character recognition according to an embodiment of the present invention. DETAILED DESCRIPTION

[0019] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present invention, and should not be construed as limiting the present invention.

[0020] The text detection method based on optical character recognition according to an embodiment of the present invention is described below with reference to the accompanying drawings.

[0021] See also Figure 1 , Figure 1 FIG. 1 is a flow chart of a text detection method based on optical character recognition according to an embodiment of the present invention. Figure 1 As shown, the text detection method based on optical character recognition includes the following steps: S101, obtaining a screen image, and performing binarization processing on the screen image to generate a binarized image corresponding to the screen image.

[0022] As an example, first, a screen image of an electronic device is captured to obtain a screen image; then, the captured screen image is binarized using the Otsu algorithm to generate a binary image corresponding to the screen image.

[0023] S102, performing single character recognition on the binary image to obtain frame coordinate information and confidence level corresponding to each character in the binary image.

[0024] S103, filtering each word based on a preset confidence threshold.

[0025] In some embodiments, each text is filtered based on a preset confidence threshold, including: determining whether the confidence corresponding to the current text is less than the preset confidence threshold; if the confidence corresponding to the current text is less than the preset confidence threshold, filtering the current text.

[0026] As an example, first, single character recognition is performed on the binary image to identify the frame coordinate information corresponding to each character and the confidence corresponding to each character; then, it is determined whether the confidence of the current character is greater than or equal to the preset confidence threshold; if not, it is determined that the confidence of the current character is low, which is an incorrectly recognized character, Chinese and English punctuation marks, or special symbols. Then, the current character with low confidence is deleted to filter out these invalid information.

[0027] S104: for each filtered character, determine whether each character is an abnormal character based on the frame coordinate information.

[0028] S105, if yes, mark the abnormal text.

[0029] In some embodiments, judging whether each text is an abnormal text based on the border coordinate information includes: calculating the actual width and actual height corresponding to the current text according to the border coordinate information; judging whether the actual width is less than a preset width threshold or whether the actual height is less than a preset height threshold; if the actual width is less than the preset width threshold or the actual height is less than the preset height threshold, determining that the current text is an abnormal text; if the actual width is greater than or equal to the preset width threshold and the actual height is greater than or equal to the preset height threshold, calculating the number of binarized pixels within the area range of the current text according to the border information; judging whether the number of binarized pixels is greater than the preset pixel number threshold; if not, determining that the current text is an abnormal text.

[0030] That is to say, first, the actual width and actual height of the current text are calculated according to the frame coordinate information; then, the actual width and actual height are determined to be too small according to the preset width threshold and the preset height threshold. If so, the current text is directly determined to be abnormal text. If not, the text area of ​​the current text is further determined according to the frame information; then, the sum of the binary pixels within the text area of ​​the current text is calculated; then, it is determined whether the value of the sum of the binary pixels is greater than the preset pixel number threshold; if not, the current text is determined to be too small and abnormal.

[0031] In some embodiments, judging whether each text is an abnormal text based on the border coordinate information includes: obtaining a preset underline thickness value; determining the underline area corresponding to the current text according to the border coordinate information and the preset underline thickness value; detecting the underline area to determine whether there is an underline in the underline area; if so, determining that the current text is an abnormal text.

[0032] In some embodiments, the underline area is detected to determine whether there is an underline in the underline area, including: determining whether there are continuous black pixels in the underline area; if so, calculating the length value of the continuous black pixels; determining whether the length value of the continuous black pixels is greater than a preset length threshold; if so, determining that there is an underline in the underline area.

[0033] As an example, first, obtain the preset underline thickness value; then, determine the underline area corresponding to the current text according to the frame coordinate information (the frame coordinate information can be the coordinates of the four edge corners of the text) and the preset underline thickness value (wherein the specific value of the preset underline thickness value can be configured according to the actual situation); then, detect the underline area of ​​the current text to determine whether there is an underline in the underline area. Specifically, first, determine whether there are continuous black pixels in the underline area; if so, calculate the length value of the continuous black pixels; then, determine whether the length value of the continuous black pixels is greater than the preset length threshold, then determine that there is an underline area in the underline area, and then determine the current text as abnormal text.

[0034] In some embodiments, for each filtered text, whether each text is an abnormal text is determined based on the border coordinate information, including: calculating the center point coordinates of the current text according to the border coordinate information, and calculating the actual width value and the actual height value of the current text according to the border coordinate information; calculating the italic angle value corresponding to the current text according to the center point coordinates, the actual width value and the actual height value; determining whether the italic angle value is greater than a preset angle value threshold; if so, determining that the current text is an abnormal text.

[0035] In some embodiments, for each text after filtering, whether each text is an abnormal text is determined based on the border coordinate information, including: determining the text area corresponding to the current text according to the border coordinate information; filtering the text area to remove white pixels in the text area to obtain a preselected pixel set; calculating the binarized average color value according to the preselected pixel set and the binarized image; calculating the average color value of the original image according to the preselected pixel set and the screen image; calculating the difference between the binarized average color value and the average color value of the original image; determining whether the difference is greater than a preset difference threshold; if so, determining that the current text is an abnormal text.

[0036] As an example, first, the border coordinate information is the coordinates of the four edge corner points corresponding to the current text; then, the text area corresponding to the current text is determined on the screen image based on the coordinates of the four edge corner points; then, filtering is performed based on the text area. Specifically, in the binary image, the white pixels in the text area are filtered out to complete the filtering of the text area and obtain a pre-selected pixel point set; then, the binary average color value of the other pixels in the text area of ​​the current text, except the white pixels, is calculated based on the pre-selected pixel point set; then, in the screen image (i.e., the original image), the original image average color value of the other pixels in the text area of ​​the current text, except the white pixels, is calculated based on the pre-selected pixel point set; then, the difference between the binary average color value and the original image average color value is calculated; then, it is determined whether the difference is greater than the preset difference threshold, and if so, the current text is determined to be abnormal text.

[0037] In summary, according to the text detection method based on optical character recognition according to an embodiment of the present invention, first, a screen image is obtained, and the screen image is binarized to generate a binarized image corresponding to the screen image; then, single text recognition is performed on the binarized image to obtain the border coordinate information and confidence corresponding to each text in the binarized image; then, each text is filtered based on a preset confidence threshold; then, for each filtered text, whether each text is an abnormal text based on the border coordinate information; then, if so, the abnormal text is marked; thereby achieving effective identification of abnormal text in the text and preventing abnormal text from being accidentally deleted and becoming unreadable.

[0038] In a second aspect, an embodiment of the present invention proposes a computer-readable storage medium on which an optical character recognition-based text detection program is stored. When the optical character recognition-based text detection program is executed by a processor, the optical character recognition-based text detection method as described above is implemented.

[0039] Thirdly, Figure 2 As shown, an embodiment of the present invention proposes a text detection system based on optical character recognition, and the text detection system based on optical character recognition includes: an acquisition module 10, a recognition module 20, a filtering module 30, a judgment module 40 and a marking module 50.

[0040] The acquisition module 10 is used to acquire a screen image and perform binarization processing on the screen image to generate a binary image corresponding to the screen image; The recognition module 20 is used to perform single-character recognition on the binary image to obtain the frame coordinate information and confidence level corresponding to each character in the binary image; The filtering module 30 is used to filter each word based on a preset confidence threshold; The judging module 40 is used to judge whether each character after filtering is an abnormal character based on the frame coordinate information; The marking module 50 is used to mark the abnormal text when determining that the text is abnormal text.

[0041] In some embodiments, each text is filtered based on a preset confidence threshold, including: determining whether the confidence corresponding to the current text is less than the preset confidence threshold; if the confidence corresponding to the current text is less than the preset confidence threshold, filtering the current text.

[0042] It should be noted that the above description of the text detection method based on optical character recognition is also applicable to the text detection system based on optical character recognition, and will not be elaborated here.

[0043] In summary, according to the text detection system based on optical character recognition in an embodiment of the present invention, an acquisition module is set to acquire a screen image, and the screen image is binarized to generate a binarized image corresponding to the screen image; the recognition module is used to perform single text recognition on the binarized image to obtain the border coordinate information and confidence corresponding to each text in the binarized image; the filtering module is used to filter each text based on a preset confidence threshold; the judgment module is used to judge whether each text after filtering is an abnormal text based on the border coordinate information; the marking module is used to mark the abnormal text when the text is judged to be an abnormal text; thereby achieving effective recognition of abnormal text in the text and preventing abnormal text from being accidentally deleted and becoming unreadable.

[0044] It should be noted that the logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be specifically implemented in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in combination with these instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in combination with these instruction execution systems, devices or apparatuses. More specific examples (non-exhaustive list) of computer-readable media include the following: an electrical connection portion with one or more wirings (electronic device), a portable computer disk box (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium and then editing, interpreting or processing in other suitable ways if necessary, and then stored in a computer memory.

[0045] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0046] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "examples", "specific examples", or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.

[0047] In the description of the present invention, it is to be understood that the terms “center”, “longitudinal”, “lateral”, “length”, “width”, “thickness”, “up”, “down”, “front”, “back”, “left”, “right”, “vertical”, “horizontal”, “top”, “bottom”, “inside”, “outside”, “clockwise”, “counterclockwise”, “axial”, “radial”, “circumferential”, etc., indicating orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the referred device or element must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as limiting the present invention.

[0048] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In the description of the present invention, the meaning of "plurality" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined.

[0049] In the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", "connected", "fixed" and the like should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; 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 the internal connection of two elements or the interaction relationship between two elements, unless otherwise clearly defined. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0050] In the present invention, unless otherwise clearly specified and limited, a first feature being "above" or "below" a second feature may mean that the first and second features are in direct contact, or the first and second features are in indirect contact through an intermediate medium. Moreover, a first feature being "above", "above" or "above" a second feature may mean that the first feature is directly above or obliquely above the second feature, or simply means that the first feature is higher in level than the second feature. A first feature being "below", "below" or "below" a second feature may mean that the first feature is directly below or obliquely below the second feature, or simply means that the first feature is lower in level than the second feature.

[0051] Although the embodiments of the present invention have been shown and described above, it is to be understood that the above embodiments are exemplary and are not to be construed as limitations of the present invention. A person skilled in the art may change, modify, replace and vary the above embodiments within the scope of the present invention.

Claims

1. A text detection method based on optical character recognition, characterized in that: The following steps are involved: Acquire a screen image, and perform binarization processing on the screen image to generate a binarized image corresponding to the screen image; Performing single character recognition on the binary image to obtain frame coordinate information and confidence level corresponding to each character in the binary image; Filter each word based on a preset confidence threshold; For each filtered character, judging whether each character is an abnormal character based on the frame coordinate information; If so, mark the unusual text.

2. The text detection method based on optical character recognition as claimed in claim 1, characterized in that: Filter each word based on a preset confidence threshold, including: Determine whether the confidence level corresponding to the current text is less than the preset confidence threshold; If the confidence level corresponding to the current text is less than the preset confidence threshold, the current text is filtered.

3. The text detection method based on optical character recognition as claimed in claim 1, characterized in that: Judging whether each character is an abnormal character based on the frame coordinate information includes: Calculate the actual width and height of the current text according to the frame coordinate information; Determine whether the actual width is less than a preset width threshold or whether the actual height is less than a preset height threshold; If the actual width is smaller than the preset width threshold or the actual height is smaller than the preset height threshold, determining that the current text is an abnormal text; If the actual width is greater than or equal to the preset width threshold and the actual height is greater than or equal to the preset height threshold, then calculating the number of binarized pixels within the area of ​​the current text according to the border information; Determine whether the number of binarized pixels is greater than a preset pixel number threshold; If not, it is determined that the current text is an abnormal text.

4. The text detection method based on optical character recognition as claimed in claim 1, characterized in that: Judging whether each character is an abnormal character based on the frame coordinate information includes: Get the preset underline thickness value; Determine the underline area corresponding to the current text according to the frame coordinate information and the preset underline thickness value; Detecting the underlined area to determine whether there is an underline in the underlined area; If so, it is determined that the current text is an abnormal text.

5. The text detection method based on optical character recognition as claimed in claim 4, characterized in that: Detecting the underlined area to determine whether there is an underline in the underlined area includes: Determine whether there are continuous black pixels in the underline area; If yes, then calculate the length value of the continuous black pixels; Determine whether the length of the continuous black pixels is greater than a preset length threshold; If yes, it is determined that there is an underline in the underline area.

6. The text detection method based on optical character recognition as claimed in claim 1, characterized in that: For each filtered character, judging whether each character is an abnormal character based on the frame coordinate information includes: Calculating the center point coordinates of the current text according to the frame coordinate information, and calculating the actual width and height of the current text according to the frame coordinate information; Calculate the italic angle value corresponding to the current text according to the center point coordinates, the actual width value and the actual height value; Determine whether the italic angle value is greater than a preset angle value threshold; If so, it is determined that the current text is an abnormal text.

7. The text detection method based on optical character recognition as claimed in claim 1, characterized in that: For each filtered character, judging whether each character is an abnormal character based on the frame coordinate information includes: Determine the text area corresponding to the current text according to the frame coordinate information; Filtering the text area to remove white pixels in the text area to obtain a preselected pixel set; Calculate the binarized average color value according to the preselected pixel point set and the binarized image; Calculate the average color value of the original image according to the preselected pixel point set and the screen image; Calculating the difference between the binarized average color value and the original image average color value; Determine whether the difference is greater than a preset difference threshold; If so, it is determined that the current text is an abnormal text.

8. A computer-readable storage medium, characterized in that: A text detection program based on optical character recognition is stored thereon, and when the text detection program based on optical character recognition is executed by a processor, the text detection method based on optical character recognition as described in any one of claims 1-7 is implemented.

9. A text detection system based on optical character recognition, characterized in that: include: An acquisition module, the acquisition module is used to acquire a screen image and perform binarization processing on the screen image to generate a binary image corresponding to the screen image; A recognition module, the recognition module is used to perform single-word recognition on the binary image to obtain frame coordinate information and confidence level corresponding to each word in the binary image; A filtering module, the filtering module is used to filter each word based on a preset confidence threshold; A judgment module, the judgment module is used to judge whether each character after filtering is an abnormal character based on the frame coordinate information; The marking module is used to mark the abnormal text when judging that the text is abnormal text.

10. The text detection system based on optical character recognition as claimed in claim 9, characterized in that: Filter each word based on a preset confidence threshold, including: Determine whether the confidence level corresponding to the current text is less than the preset confidence threshold; If the confidence level corresponding to the current text is less than the preset confidence threshold, the current text is filtered.

Images