JP2000020714A - Image processing method, apparatus and recording medium storing image processing function - Google Patents

Abstract

(57) [Summary] [Problem] To provide a binary image free of noise that hinders recognition of a background image. A grayscale image of a document to be binarized is input (S1) and stored (S2). Next, attention is paid to one pixel, and it is determined whether or not the pixel corresponds to an edge of a character or a ruled line (S3). For the pixel determined to be an edge, the pixel value in the binary image is determined (S4) and stored. These operations are performed for all pixels on the original image (S
5) Find all connected components of non-edge pixels (S
6). Attention is paid to the pixels which are in contact with the periphery of a certain connected component and for which the pixel value has already been determined, and the number of black pixels and white pixels among them is counted (S7). These are compared (S8), and if there are more black pixels, the entire connected component is registered as a black pixel (S10). Otherwise, it is registered as a white pixel (S9). This operation is performed for all connected components (S11), and finally, a binary image is generated (S12) and output (S13).

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing method, an apparatus, and a recording medium storing an image processing function, and more particularly, to a method for optically recognizing characters written on a sheet of paper. The present invention relates to an image processing method for generating a binary image suitable for recognition and preprocessing thereof.

[0002]

2. Description of the Related Art Most of the current optical character recognition technologies recognize binary images, and the accuracy is greatly affected by the image quality. Therefore, obtaining an appropriate binary image is indispensable for recognition. That is what. Conventionally,
Many methods have been devised for obtaining an appropriate binary image from a grayscale image. For example, Japanese Unexamined Patent Publication No. Hei 6-18439 discloses an edge component extracted from an image signal, and identifies a character area, a gradation area, and an area having intermediate properties based on the strength of the edge component. By binarizing only the image using the dither method, a high-quality binary image is obtained.

[0003]

However, the method disclosed in Japanese Patent Publication No. 6-18439 does not aim at character recognition, and thus has a corresponding drawback. For example, the background is not originally a target of character recognition, but if an intermediate color is used as the background, this area is subjected to gradation processing and is expressed in dither. In this case, since there are many black pixels, the area is a character area. May not be distinguished, and the recognition process may be forcibly applied. In the case of character recognition, it is desirable that the background, which is an intermediate color on the original image, does not express the intermediate color on the binary image, and forcibly expressing it will generate noise that hinders recognition. I can only do that.

The present invention has been made in consideration of the above-described circumstances, and utilizes the fact that the ruled lines and characters have different densities from the background color in the vicinity. Provided is an image processing method and apparatus capable of providing a binary image free from noise that hinders recognition in the background by clearly expressing characters and ruled lines, and a recording medium storing the image processing function. It is done for the purpose of.

[0005]

According to the first aspect of the present invention, it is determined whether or not each pixel in an input grayscale image is an edge, and the pixels determined as edges are binarized.
For pixels not determined as edges, by performing a binarization by examining a contact portion between the connected components of the pixels not determined as edges and the pixels determined as the edges,
It is characterized in that a binary image is generated from an input grayscale image.

According to a second aspect of the present invention, in the first aspect of the present invention, for a pixel not determined as the edge, a connected component of the pixel determined as the edge is extracted, and a circumscribing of the extracted connected component is performed. Among the pixels in the rectangle, binarization is performed on pixels that are not determined to be edges, and a contact portion between a connected component of a pixel that has not been binarized in the above processing and a pixel that has already been binarized is examined. In this way, binarization of the pixels that have not been binarized is performed.

According to a third aspect of the present invention, there is provided an edge determining means for determining whether or not each pixel in an input grayscale image is an edge, and binarization of the pixel determined as an edge by the edge determining means. First binarizing means for performing, and examining a contact portion between a connected component of a pixel not determined as an edge by the edge determining means and a pixel determined as the edge, a pixel of the pixel not determined as the edge. And a second binarizing means for performing binarization, wherein a binary image is generated from the input grayscale image.

According to a fourth aspect of the present invention, in the third aspect of the present invention, there is provided a connected component extracting means for extracting a connected component of the pixel determined to be the edge, and A third binarization unit for binarizing a pixel not determined as an edge, wherein the second binarization unit performs binarization by the first and third binarization units. The non-binarized pixel is binarized by examining a contact portion between a connected component of a pixel that has not been binarized and a pixel that has already been binarized.

According to a fifth aspect of the present invention, there is provided an edge determining function for determining whether each pixel in an input grayscale image is an edge, and a first binary function for performing binarization on a pixel determined as an edge. A binarization function, and for a pixel not determined as an edge, a second binary value for performing binarization by examining a contact portion between a connected component of the pixel not determined as the edge and the pixel determined as an edge And a computer program for realizing a function of generating a binary image from an input grayscale image by a computerization function.

According to a sixth aspect of the present invention, in the fifth aspect of the present invention, a connected component extracting function for extracting a connected component of the pixel determined to be the edge, and And a third binarizing function for binarizing a pixel that is not determined to be an edge, wherein the second binarizing function is performed by the first and third binarizing functions. The non-binarized pixel is binarized by examining a contact portion between a connected component of the non-binarized pixel and a pixel that has already been binarized.

[0011]

FIG. 1 is a block diagram for explaining an embodiment of a recording medium storing an image processing method, an apparatus and an image processing function according to the present invention. Processing target gray image input means for obtaining an image, reference numeral 2 denotes a processing target gray image storage means for storing a gray image, and reference numeral 3 denotes whether or not each pixel is an edge of a character or a ruled line from the stored gray image. A character edge discriminating means, 4 is an edge pixel binarizing means for determining a binarized pixel value of a pixel determined as a character edge, and 5 is a binarization result storage for temporarily storing a binarization result. Means 6 is a non-edge pixel connected component extracting means for examining connectivity of pixels not determined to be edges and extracting connected ones, and 7 is already binarized around the non-edge pixel connected components. A peripheral pixel counter that counts the number of pixels for each black / white , 8 binary image generating means for generating a binary image on the basis of the information stored in the binarization result storage means, 9 is a binary image output means for outputting the binary image obtained.

FIG. 2 is a flowchart of the processing of the embodiment shown in FIG. 1, and FIG. 3 is a diagram for explaining an image processed by the embodiment shown in FIG. Hereinafter, the processing operation will be described with reference to the flowchart shown in FIG. A gray-scale image of a document to be binarized is input by the gray-scale image input means 1 such as a scanner (S1), and stored in the gray-scale image storage means 2 (S2). next,
The character edge discriminating means 3 pays attention to a certain pixel and determines whether this pixel corresponds to the edge of a character or a ruled line (S3). As a method of determining whether or not an edge is an edge, a method of examining pixel values in a range of several pixels square around the target pixel, and determining an edge when a difference between the maximum value and the minimum value of the pixel value is equal to or greater than a certain threshold is used There is.

For the pixel determined as an edge, the pixel value in the binary image is determined by the edge pixel binarizing means 4 (S4) and stored in the binarization result storing means 5. As a method, the average value of the maximum value and the minimum value of the pixel values in a range of several pixels around the target pixel is set as a threshold, and if it is closer to white than this threshold, it is determined as a white pixel, otherwise it is determined as a black pixel There are methods.

The image shown in FIG. 3B is obtained by performing the above-described operation on the grayscale image shown in FIG. 3A. Of the edges, a portion close to a high-density region is black, and a portion close to a low-density region is a white pixel. In the example shown in FIG. 3B, among the pixels determined to be edges, the pixels determined to be black / white pixels are represented by black / white, respectively. Undetermined pixels are shown in gray.

These operations are performed for all the pixels on the original image (S5). Next, referring to the binarization result storage means 5, the non-edge pixel connected component extracting means 6 connects the non-edge pixel connected components. (S6). FIG.
In the example shown in (B), a connected component of gray pixels is obtained. Then, attention is paid to the pixels which have already determined the pixel value and which are in contact with the periphery of a certain connected component, and the peripheral pixel counting means 7 counts how many black pixels and white pixels are present among them (S7).

These are compared (S8). If there are more black pixels, the entire connected component is registered as a black pixel in the binarization result storage means 5 (S10). Otherwise, it is registered as a white pixel (S9). This operation is performed for all connected components (S11). Finally, a binary image is generated by the binary image generation means 8 based on the contents of the binarization result storage means 5 (S12), and output by the binary image output means 9 (S13).

In the embodiment shown in FIGS. 1 and 2, in the case of a character having a small size, the edge inside the character cannot be sufficiently extracted, and the inside of the character cannot be well represented on a binary image. Sometimes. Therefore, at the stage when the edge extraction is completed, the connected components of the edge pixels are extracted, and binarization with higher reproducibility is performed for each character. By binarizing non-edge pixels as in the example, the inside of a small character can be expressed.

FIG. 4 is a block diagram for explaining another embodiment of a recording medium storing an image processing method and apparatus and an image processing function according to the present invention. In FIG. 12 is a connected component circumscribed rectangle determining means, 13 is an undecided in-rectangle pixel binarizing means, and the other parts having the same operation as in FIG. 1 are denoted by the same reference numerals as those in the embodiment shown in FIG.

FIG. 5 is a flowchart of the processing of the embodiment shown in FIG. The configuration of the embodiment shown in FIG.
4 is added to the embodiment shown in FIG. 4. Similarly, the embodiment shown in FIG. 5 is obtained by adding the portion B shown in FIG. 5 to the embodiment shown in FIG. It is a thing.
Hereinafter, the processing operation will be described with reference to the flowchart shown in FIG.

After the binarization of the edge portion is completed in step S5, in step S21, the connected component of the black pixel portion is extracted by the black pixel connected component extracting means 11, and then the connected component circumscribed rectangle determining means 12 is extracted. To determine the range of the circumscribed rectangle (S22). If this rectangle constitutes a character, there is a possibility that a fine structure is included in the character, and there is a possibility that edge extraction has not been completed to that extent. However, since it is difficult to determine whether a character is a character, here, the size of the character is determined in advance, and a rectangle that is supposed to constitute the character is determined based on the specified value of the rectangle size range according to that ( S23).

Here, there is no adverse effect even if a non-character is determined to be a character. Conversely, if it is determined that what should be a character is not a character, there is a possibility that the internal structure of the character can not be reproduced, but this often occurs when the character is too large I do. in this case,
This is not a problem since the edge extraction is often performed inside.

The rectangle determined to be a character is binarized by the undecided pixel binarizing means 13 only for undecided pixels within the rectangle (S24). There are various methods of this binarization, but the pixel value of the pixel in the rectangle in the original image is checked, and the average value of the maximum value and the minimum value is obtained as a threshold,
There is a method of determining a white pixel if the pixel is closer to white than this threshold, and a black pixel otherwise if the pixel is closer to white. The steps up to this point are performed for all connected components (S25), and when the processing is completed, the flow proceeds to the processing of step S6.

FIG. 6 is a diagram for explaining an example in which the processing according to the embodiment shown in FIGS. 4 and 5 has been performed. FIG.
FIG. 6B is an example of a result of performing up to binarization in character units.
Is an example of the final result of binarization.

FIG. 7 is a block diagram for explaining another embodiment of a recording medium storing an image processing method and apparatus and an image processing function according to the present invention. In FIG. 7, reference numeral 21 denotes a CPU; Is a hard disk, 24 is an input device, 25
Is a CD-ROM drive, 26 is a display, and 27 is a recording medium.

The embodiment shown in FIG. 7 is a general-purpose processing device comprising a mouse (not shown) and the like in addition to the above-described components. A program for realizing an identification processing function and a processing procedure is recorded. A document image of a form to be registered / processed is stored, for example, on a hard disk 23 or the like. A program for implementing a processing function or procedure is read and executed, and the binarized result is output to the display 26 or the like.

[0026]

As is apparent from the above description, according to the present invention, by utilizing the fact that the ruled lines and characters have different densities from the neighboring background colors, even if an intermediate color is used for the background. Since the characters and the ruled lines are clearly expressed, it is possible to provide a binary image free of noise which also hinders the recognition of the background.

[Brief description of the drawings]

FIG. 1 is a configuration diagram illustrating an embodiment of a recording medium storing an image processing method, an apparatus, and an image processing function according to the present invention.

FIG. 2 is a flowchart of a process of the embodiment shown in FIG. 1;

FIG. 3 is a diagram for explaining an image processed by the embodiment shown in FIG. 2;

FIG. 4 is a configuration diagram for explaining another embodiment of a recording medium storing an image processing method and apparatus and an image processing function according to the present invention.

FIG. 5 is a flowchart of a process of the embodiment shown in FIG. 4;

FIG. 6 is a diagram for explaining an image processed by the embodiment shown in FIGS. 4 and 5;

FIG. 7 is a configuration diagram for explaining another embodiment of a recording medium storing an image processing method and apparatus and an image processing function according to the present invention.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Gray-scale image input means for processing, 2 ... Gray-scale image storage means for processing, 3 ... Character edge discriminating means, 4 ... Edge pixel binarization means, 5 ... Binarization result storage means, 6 ... Non-edge pixel connection component Extraction means 7, peripheral pixel counting means 8, binary image generation means 9, binary image output means 11, black pixel connected component extracting means 12, connected component circumscribed rectangle determining means 13,
... Undecided pixel binarization means in a rectangle, 21 ... CPU, 22 ...
Memory 23 hard disk 24 input device 25
... CD-ROM drive, 26 ... Display, 27 ...
recoding media.

Claims (6)

[Claims] 1. A method for determining whether each pixel in an input grayscale image is an edge, performing binarization on a pixel determined as an edge, and performing edge conversion on a pixel not determined as an edge. An image characterized by generating a binary image from an input grayscale image by performing a binarization by examining a contact portion between a connected component of a pixel not determined as the pixel and the pixel determined as the edge. Processing method. 2. The image processing method according to claim 1, wherein
For the pixels not determined as the edges, the connected components of the pixels determined as the edges are extracted, and among the pixels in the circumscribed rectangle of the extracted connected components, the binary values of the pixels not determined as the edges are extracted. The binarization of the non-binarized pixel is performed by examining the contact portion between the connected component of the pixel that has not been binarized in the above process and the pixel that has already been binarized. An image processing method comprising: 3. An edge determining means for determining whether each pixel in the input grayscale image is an edge, and a first binarizing means for binarizing the pixel determined as an edge by the edge determining means. A binarizing unit that binarizes the pixel that is not determined as the edge by examining a contact part between the connected component of the pixel that is not determined as the edge and the pixel that is determined as the edge by the edge determining unit; An image processing apparatus, comprising: a second binarization unit, which generates a binary image from an input grayscale image. 4. The image processing apparatus according to claim 3, wherein: a connected component extracting unit that extracts a connected component of the pixel determined to be the edge; and a pixel within a circumscribed rectangle of the extracted connected component. A third binarization unit for performing binarization of a pixel not determined as an edge, wherein the second binarization unit is binarized by the first and third binarization units. An image processing apparatus for performing binarization of a pixel that has not been binarized by checking a contact portion between a connected component of a pixel that has not been binarized and a pixel that has already been binarized. 5. An edge determination function for determining whether each pixel in an input gray-scale image is an edge, a first binarization function for performing binarization on a pixel determined to be an edge, For a pixel not determined as an edge, a second binarization function of performing a binarization by examining a contact portion between a connected component of the pixel not determined as the edge and the pixel determined as an edge, A recording medium storing an image processing function characterized by recording a program for causing a computer to realize a function of generating a binary image from an input grayscale image. 6. A recording medium storing the image processing function according to claim 5, wherein a connected component extracting function for extracting a connected component of the pixel determined to be the edge, and a circumscribing rectangle of the extracted connected component. A third binarization function of performing binarization of pixels that are not determined to be edges among pixels of
The second binarization function checks the contact portion between a connected component of a pixel that has not been binarized by the first and third binarization functions and a pixel that has already been binarized. A recording medium storing an image processing function characterized by performing binarization of a pixel that has not been binarized.