JP3127676B2 - Image processing apparatus and image processing method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing apparatus and an image processing method for binarizing a grayscale image including a character image into a character portion and a background portion.

[0002]

2. Description of the Related Art Conventionally, an image processing apparatus determines a binarization threshold for image data of a document or the like based on density information in the image data, and uses a single threshold for a black portion (character portion) and a white portion. (A background part), a method of generating a binary image, and a method of dividing a document image into an area including a character and an area such as a photograph and determining a threshold value for each of an area such as a photograph and an area including a character and determining each area. In many cases, a method of generating a binary image is used.

[0003]

However, in the above-described configuration, in the case where unevenness in shading occurs in the image with respect to the character area, or when the density of the character is non-uniform, depending on the character in the binarized image, There is a problem that the lines are crushed or blurred. Further, in a region where the background density is different from the entire background density, there is a problem that the entire region becomes a black portion and a character portion cannot be extracted.

In view of the above problems, the present invention binarizes the entire area of a character image using a single threshold, extracts an area for each individual character, and sets an optimum binarization threshold for each character. By generating a binary image again, the purpose is to generate a good binary image with less collapse and blurring of character lines even when shading unevenness occurs in the image or when the density of characters is uneven. And Further, even when the density of the background portion varies depending on the location in the image, the character portion can be binarized as a black portion and the background portion as a white portion.

[0005]

In order to solve the above-mentioned problems, an image processing method according to the present invention specifies an image area to be binarized for gray image data including a character image such as a document .
Area specifying step, and within the area for the specified area
Binarization threshold determination that determines a binarization threshold based on density information
Setting step and the area specifying step
The image area determined in the binarization threshold determination step is determined.
Binarization using the set binarization threshold
Generating a binary image, and generating the binary image.
The image area generated by the forming step is a one-character area.
Area determination step of determining whether or not the
If it is determined that the area is not one character area
In this case, the binary image is divided into regions where black portions exist,
The area to pass each area divided in the area specifying step
A dividing step and the one-character area determining step,
The areas determined to be one character area are combined to form the final two
And a binary image synthesizing step as a value image .

Further , the image processing method of the present invention is characterized in that the binary
Thresholding step is performed for the specified binarization area.
Binary using both density information in area and area size information
It is provided with a configuration for determining the activation threshold .

[0007] The image processing apparatus of the present invention can be used for character images such as sentences.
Image to be converted to binary image data
An image processing apparatus for performing an image processing method, wherein the image
The processing device is used for processing grayscale image data including character images such as documents.
Image based on the binarization threshold for the specified area
And a binary image generation unit for generating a plurality of specified regions.
A binary image synthesizing unit for synthesizing a binary image, wherein the image
The processing method is applied to grayscale image data including character images such as documents.
On the other hand, an area specifying step for specifying an image area to be binarized
And the specified area based on the density information in the area
A binarization threshold determining step of determining a binarization threshold,
The image area specified in the area specifying step is
Binarization threshold determined in the binarization threshold determination step
And binarize the binary image of the image area with the binary image
Generating a binary image generated by the component,
The image area generated by the image generation step is one character area
An area determining step of determining whether the area is an area,
In the area determination step, it is determined that the area is not a one-character area.
The binary image is divided into regions where black portions exist.
And pass each divided area to the area designation step
An area dividing step and the one-character area determining step.
Therefore, each area determined to be one character area is designated,
The final binary image is synthesized by the binary image synthesizing unit.
And a binary image synthesizing step.
It is a thing. Further, the image processing apparatus of the present invention is characterized in that
The thresholding step for determining the binarization threshold is performed for the specified binarization area.
Using both the intra-area density information and the area size information
It is provided with a configuration for determining a valuation threshold.

[0008]

According to the present invention, an optimum binarization threshold value is obtained for each character area extracted from a grayscale image by the above-described structure.
By binarizing each area with an individual threshold value obtained for each character area and creating a binary image, a high-quality binary image with little collapse or blurring of character lines can be obtained. Further, even if the density of the background portion differs depending on the location by repeating the operation of dividing the binary image into regions where black portions are present, obtaining a binarization threshold for each region and generating a binary image again, The binarization can be performed so that the background portion becomes a white portion without fail. Further, when determining the binarization threshold, by using both the area size and the density information of the area to be binarized, the binarization threshold can be set finely,
It is possible to prevent a single character from being divided into a plurality of areas due to blurring of a character line or the like based on a binarization threshold determined in an initial stage, and to obtain a high-quality binary image.

[0009]

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram of an image processing apparatus according to a first embodiment of the present invention. In FIG. 1, reference numeral 11 denotes a binarization threshold value determination unit which determines a binarization threshold value for a designated region in the grayscale image data 16 including a character image based on the in-region density information. Reference numeral 12 denotes an entire image generation unit, which designates the entire gray-scale image area for the binarization threshold determination unit 11 and binarizes the whole with a single threshold determined by the binarization threshold determination unit 11 to convert the binary image Generate. Reference numeral 13 denotes a character area extracting unit that extracts a character image area for each character from the binary image generated by the entire binary image generating unit 12. 1
Reference numeral 4 denotes a partial binary image generation unit which designates each one-character image region in the grayscale image data extracted by the character region extraction unit 13 to the binarization threshold value determination unit 11;
In step 1, a partial binary image of each region is generated with each binarization threshold determined for each region. Reference numeral 15 denotes a binary image synthesizing unit which synthesizes the binary images of the respective regions generated by the partial binary image generating unit 14 to reconstruct the entire binary image.

The operation of the image processing apparatus configured as described above will be described below with reference to FIGS. 1, 2, 3, 4, and 5. FIG. 2 shows the input gray-scale image data.
Reference numerals 2, 23, and 24 indicate respective character areas in the image. In the grayscale image, each pixel is represented by a density value, and the higher the density value, the closer to white and the lower the value, the closer to black.

First, the entire binary image generation unit 12 specifies the entire gray-scale image data area 21 input to the binarization threshold value determination unit 11. The binarization threshold determination unit 11 determines a binarization threshold based on the grayscale information of the specified entire area. For example, a method of counting a distribution (number of pixels) for each density in a designated area and setting a threshold value such that a black portion becomes a constant ratio of the entire area, a method of setting the number of pixels of a black portion to N1, When the average of the density values is m1, the number of pixels in a white portion is N2, and the average of the density values is m2, N1 × N2 × (m1-m2)
A method of determining a threshold value that maximizes the variance of the average value of the white portion and the black portion represented by × (m1−m2) as a binarization threshold value (discrimination analysis method), or preparing a plurality of threshold values in advance There is a method of selecting a threshold value according to the average density in the designated area, and the like. In the present embodiment, it is assumed that the image is selected from a plurality of thresholds prepared in advance by averaging the image density values. For example, when the density value of each pixel in the grayscale image data 21 is given at 256 levels from 0 (black) to 255 (white), the binarization threshold determination unit 11 sets the average density value in the area to 170 If it is less than or equal to the threshold 170, if it is 171 to 180, it is the threshold 180, if it is 181-190, it is the threshold 190, 19
0 to 210, threshold value 200, 210 or more, 220
Shall be determined. For example, if the average density of the entire area 21 specified by the entire binary image generation unit 12 is 2
If it is 05, the threshold value 200 is selected, and the threshold value 200 is passed to the entire binary image generation unit 12. The entire binary image generation unit 12 uses the threshold value determined by the binarization threshold value determination unit 11 to define pixels below the threshold value in the black portion (character portion) and pixels above the threshold value in the white portion in the entire grayscale image data area 21. (Background portion), and the entire binary image 31 is generated.

FIG. 3 shows a binary image generated by the entire binary image generating unit, where 31 is the entire image data, and 32, 33 and 34 are one characters extracted by the character area extracting unit 13. It shows an example of a region. From the binary image generated by the overall binary image generation unit 12, each character region is extracted by a character region extraction unit 13. For example, the image is scanned in the horizontal direction with respect to the whole binary image 31 to detect a character line by detecting a portion where a black pixel exists, and further, the image is scanned in the vertical direction for each detected character line and a black pixel is scanned. Is detected, and an area for each character is detected. For example, FIG. 4A shows the result of scanning the entire binary image in the horizontal direction for each line.
Reference numerals 2 and 43 indicate portions where black pixels exist.
FIG. 4B shows the result of scanning the character lines 41 and 43 in the vertical direction. 44, 45, and 46 are examples of the detected character regions. The partial binary image generation unit 14 specifies each one-character area extracted by the character area extraction unit 13 to the binarization threshold determination unit 11, and the binarization threshold determination unit 11 The threshold of each area is individually determined based on the density information of the area. For example, 45, 4 in FIG.
If the average density of each one-character area shown in FIGS. 6 and 47 is 182, 189, and 220, respectively, the threshold value of each area is set to 190, 1
90 and 220 are determined. In accordance with the determined threshold value, the partial binary image generation unit 14 binarizes each one-character area and generates a partial binary image as shown by 52, 53, and 53 in FIG. The partial binary image generated by the partial binary image generator 14 is reconstructed by the binary image synthesizer 15 to generate the entire binary image. For example, the entire binary image 31 is obtained by reconstructing a partial binary image of each one-character area to obtain a final binary image, 51 in FIG.

FIG. 6 shows image processing according to the second embodiment of the present invention.
FIG. 2 is a diagram illustrating a configuration of an image processing method in the apparatus .
The image processing device includes a binary image generation unit and a binary
It has an image synthesis unit. In FIG. 6, reference numeral 61 denotes an area specifying step, which specifies an area to be binarized in the gray image data 67. Reference numeral 62 denotes a binarization threshold determination step for determining a binarization threshold for a designated area in the grayscale image data 67 based on the in-area density information. Reference numeral 63 denotes a binary image generation step which is a binary image generation unit, which binarizes the area of the grayscale image specified in the area specification step 61 with the binary threshold determined in the binary threshold determination step 62. Generate a binary image. 64 is a region determination step,
It is determined whether or not the area generated in step 63, which is the binary image generation unit , is a one-character area. An area dividing step 65 divides a binary image corresponding to an area determined to be not a one-character area in the area determining step 64 for each area where black portions are continuous. 66 binary image synthesizing unit der
The binary image of each area determined to be a one-character area in the area determination step 64 is synthesized to generate a final binary image 68.

The operation of the image processing method configured as described above will be described below with reference to FIGS. 6, 7, 8, and 9. FIG. 7 shows the input grayscale image data, 71 indicates the entire grayscale image data, and 72 indicates a background portion having different density values. First, the entire image area of the grayscale image 71 is specified in the area specifying step 61 and is passed to the binarization threshold value determining step 62. In the binarization threshold determination step 62, the binarization threshold is determined based on the grayscale information in the grayscale image 71 as described in the first embodiment. For example, if the overall average density of 71 of the grayscale image is 205 ,
Der If the density value of the background portion 72 which is a portion of which is 190
What is also of gray-scale image 71 by binarization threshold determination step 62
The threshold value 200 is determined based on the average density, and a binary image 81 is generated by binarizing pixels below the threshold into a black portion and pixels above the threshold into a white portion in a binary image generation step. FIG. 8 shows the entire binary image generated in step 63 which is a binary image generating unit , where 81 is the entire image data, and 82 is an area 72 having a background portion with different density values.
5 shows an example of the binarization result of the above . In the binary image generating unit
An area determination step 64 determines whether the binary image generated in a certain step 63 is a one-character area. For example, it is determined whether or not the binarized area is a one-character area based on whether or not the binarized area has a certain size or more. If it is determined that the binary image data area 81 is not a one-character area, the area is divided into areas in which black portions continue by an area dividing step 65. For example, by scanning the binary image area in the horizontal direction as shown in the first embodiment, the continuous area 91 of the black portion shown in FIG.
It is divided into 92, 93 and 94. By further scanning in the horizontal direction, for example, the region 93 is divided into the regions shown in FIG. Each divided region is, the binary image generating unit der determined binarization threshold by binarization threshold determination step 62 for each of the regions is passed back to area designation step
In step 63, a binary image of each area is generated. For example, when the average density value of the 95 parts of the area in FIG. 9 is 160, the threshold value becomes 170, the background part 72 is binarized to white, and the binary image 96 of the 95 parts of the area shown in FIG. Is done. The binary image 96 is 1
It is determined whether or not the area is a character area, and the area is further divided into the areas shown in FIG. Each area determined to be a one-character area by the area determination step 65 is synthesized by a binary image synthesizing step 66 to generate a final binary image 68. In the second embodiment, in the binarization threshold determination step 62, a plurality of thresholds are prepared and the binarization threshold is selected and determined based on the average value of the density of the designated area. A threshold corresponding to the density value may be prepared, and the binarization threshold may be determined based on the area size and the average density in the area.
For example, a higher threshold can be prepared when the area size is larger than one character area, and a lower threshold can be prepared when the area size is one character area.

[0016]

According to the present invention as described above, according to the present invention, determine the optimum binarization threshold for each character region extracted from the concentrated light in the image, each region with a particular threshold determined for each region A binary image can be created by binarization, and a high-quality binary image with little collapse and blurring of character lines can be obtained.

Further, according to the present invention, the operation of dividing the binary image into regions where black portions are present, obtaining a binarization threshold value for each region, and generating a binary image again is repeated, thereby obtaining a background image. Even when the density of a part differs from place to place, binarization can be performed so that the background part is a white part without fail, and an optimal binarization threshold value is determined for each character area, and individual values determined for each area are determined. Each area is binarized with the threshold value of, and a binary image can be created, so that a high-quality binary image with little collapse or blurring of character lines can be obtained.

Further , according to the present invention , in the binarization threshold value determining step , the binarization threshold value is determined for the designated binarization region using both the intra-region density information and the region size information. By providing the binarization threshold determination step, the binarization threshold can be set finely, and it is possible to prevent a character from being divided into a plurality of regions due to blurring of a character line due to the binarization threshold determined in the initial stage. And a high quality binary image can be obtained.

[Brief description of the drawings]

FIG. 1 is a configuration diagram of an image processing apparatus according to a first embodiment of the present invention.

FIG. 2 is a photograph of an input image according to the first embodiment of the present invention.

FIG. 3 is a diagram showing an image showing an overall binarization result according to the first embodiment of the present invention.

FIG. 4 is a diagram showing a character area extraction result according to the first embodiment of the present invention;

FIG. 5 is a diagram showing an image showing a final binarization result in the first embodiment of the present invention.

FIG. 6 is a configuration diagram of an image processing method according to a second embodiment of the present invention.

FIG. 7 is a photograph showing an input image in Embodiment 2 of the present invention.

FIG. 8 is a diagram showing an image showing an overall binarization result according to the second embodiment of the present invention.

FIG. 9 is a diagram showing a result of dividing an image area into regions according to the second embodiment of the present invention.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 11 Binarization threshold value determination part 12 Whole binary image generation part 13 Character area extraction part 14 Partial position image generation part 15 Binary image synthesis part 16 Gray image data 17 Binary image 21 Total gray image data 22 Example of one character area 23 Example of one character area 24 Example of one character area 31 Whole binary image 32 Example of one character area 33 Example of one character area 34 Example of one character area 41 Detected character line 42 Detected character line 43 Detected Detected character line 44 Detected character line 45 Detected one-character area example 46 Detected one-character area example 47 Detected one-character area example 51 Final binarized image 52 Partial binary image example 53 Example of partial binary image 54 Example of partial binary image 61 Area designation step 62 Binarization threshold determination step 63 Binary image generation step 64 Area determination step 65 Area division step Step 66 Binary image synthesizing step 67 Gray-scale image data 71 Total gray-scale image data 72 Background part with different density values 81 Whole binary image 82 Binary result of background part with different density values 91 Horizontally divided area 92 Horizontal Region divided in the direction 93 region divided in the horizontal direction 94 region divided in the horizontal direction 95 region divided in the vertical direction 96 binarization result of the background portion having different density values

────────────────────────────────────────────────── ─── Continued on the front page (72) Inventor Satoshi Emura 1006 Kazuma Kadoma, Kadoma City, Osaka Inside Matsushita Electric Industrial Co., Ltd. (56) References JP-A-4-268988 (JP, A) (58) Survey Field (Int.Cl. ⁷ , DB name) H04N 1/403 G06K 9/38

Claims (4)

(57) [Claims] 1. An area specifying step of specifying an image area to be binarized for grayscale image data including a character image such as a document, and a binarization threshold based on density information in the area for the specified area. A binarization threshold value determining step of determining the image area specified in the area specifying step,
A binary image generation step of binarizing with the binarization threshold determined in the binarization threshold determination step and generating a binary image of the image area, and an image area generated by the binary image generation step An area determining step of determining whether the area is a one-character area; and if the area determining step determines that the area is not a one-character area, the binary image is divided for each area where a black portion exists, and the area specification is performed. a region dividing step of passing the respective regions divided into the step, by said one character area determination step, each region is determined to be a single character area combined to the binary image synthesis step of the final binary image image processing method characterized by comprising. Wherein said binarization threshold determination step, claims and determines the binarization threshold using both the area density information and area size information for a given binarized region Item 10. The image processing method according to Item 1 . 3. Gray-scale image data including a character image such as a sentence
To execute the image processing method to convert to binary image data
An image processing apparatus, wherein the image processing apparatus includes a designated image data in gray image data including a character image such as a document.
Generate a binary image based on a binarization threshold for a region
A binary image generating unit for combining binary images of a plurality of designated areas with a binary image generating unit;
And an image processing method , wherein the image processing method binarizes grayscale image data including a character image such as a document.
Area specifying step for specifying an image area to be processed, and binary processing based on the density information in the area for the specified area.
Binarization threshold value determining step of determining a binarization threshold, and the image region designated in the region designation step,
Binarization determined in the binarization threshold determination step
Binarize with a threshold, and convert the binary image of the image area to the binary image
A binary image generating step generated by an image generating unit, and an image area generated by the binary image generating step is
Area determination step for determining whether the area is a one-character area
The area determination step determines that the area is not a one-character area.
When the binary image is cut off, the binary image is
And each area divided in the area specifying step is
In the area dividing step to be passed and the one-character area determining step, the one-character area is determined.
Each area determined to be
Image compositing step for compositing to make the final binary image
An image processing apparatus, comprising: 4. The method according to claim 1, wherein the step of determining a binarization threshold value comprises
Density information and area size for the selected binarized area
Determining the binarization threshold using both of the information
The image processing apparatus according to claim 3, wherein: