JP2005072858A - Image processing apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an image processing apparatus capable of accurately reproducing faint letters in the case of binarizing an original image including the light letters. <P>SOLUTION: The image processing apparatus is provided with: a character region extract section 11 for extracting a character region from an original image; a filter processing section 16 for applying smoothing processing to the extracted character region; a multi-value processing section 12 for converting the character region subjected to smoothing processing into an m-value processed image; a threshold value determining section 13 for discriminating whether or not a prescribed number (e.g., one) of pixels configuring the character region exists with respect to each of m classes and determining a threshold on the basis of a result of the discrimination; a binary processing section 14 for binarizing the character region on the basis of the determined threshold; and an output section 15 for generating a print image on the basis of the binarized character region and printing out the image on recording paper. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to an image processing apparatus that binarizes a document image.

  In an image processing apparatus that binarizes an image of a document read by a document reading unit and prints it on a recording sheet, a predetermined threshold value is compared with the luminance (density) of each pixel to determine white or black. Done and printed on recording paper. For this reason, there is a problem that an image of a document including characters written with a pencil or thin characters such as vermilion cannot be binarized accurately unless the threshold value is set well.

Therefore, a technique for creating a histogram of a document image and determining a threshold based on the histogram is disclosed (Patent Document 1).
JP-A-10-65915

  However, in the method of Patent Document 1, when there are few thin characters with respect to the printed characters, the threshold is set high, and there is a possibility that the thin characters may not be reproduced.

  The present invention has been made to solve the above problems, and an object of the present invention is to provide an image processing apparatus capable of accurately reproducing a thin character when a document image including the thin character is binarized. And

  In order to solve the above-described problem, an image processing apparatus according to the present invention is an image processing apparatus that binarizes an original image represented by n (n ≧ 4) gradations. Based on the luminance of the pixel, the character area extracting means for extracting a character area composed of pixels representing the character, and the extracted character area is divided into m (m ≧ 3 and m <n). Multi-value conversion means for converting into a certain m-valued image and whether or not a predetermined number of each pixel of the m-valued image exists for each gradation, and based on the determination result, the character region Threshold value determining means for determining a threshold value for binarizing the character area, and binarizing means for binarizing the character area based on the determined threshold value.

  The image processing apparatus may further include a smoothing unit that smoothes the character region using a smoothing filter, and the multi-value conversion unit may convert the character region smoothed by the smoothing unit into an m-valued image. preferable.

  Further, the image processing apparatus further includes an edge extraction unit that performs edge extraction processing on the character region using an edge extraction filter, and the multi-value conversion unit converts the character region extracted by the edge extraction unit into an m-valued image. It is preferable to convert.

  According to the first aspect of the present invention, an original image represented by n gradations is multi-valued into an m-valued image having m (m <n) gradations, and each pixel of the m-valued image is assigned to each floor. It is determined whether or not there is a predetermined number for the key, and based on the determination result, a threshold is determined and the character area is binarized. If a character area containing thin characters such as a pencil and printed characters printed with ink is multi-valued to m gradations, any of the multi-valued images is displayed at gradations above the minimum gradation of the thin characters. Therefore, if the gradation of the boundary between the character portion of the document and the background portion of the document is clearly understood and binarized using the gradation as a threshold value, the light character Can be extracted. As a result, it is possible to accurately reproduce thin characters even when the frequency of thin characters included in a document is low compared to the case where the threshold value is determined based on a histogram as in the past. Furthermore, since it is not necessary to create a histogram, the processing can be simplified.

  According to the second aspect of the present invention, since the character area is smoothed using a smoothing filter such as a median filter or an averaging filter and then multivalued, the character area is erroneously affected by noise. Therefore, it is possible to prevent multi-values. As a result, thin characters can be reproduced more accurately.

  According to the third aspect of the present invention, since the character area is subjected to edge extraction processing by an edge extraction filter such as a differential filter and then multivalued, the character area included in the character area is accurately increased. Can be priced.

  FIG. 1 is a schematic side view mainly showing a mechanical configuration of the copying machine when the image processing apparatus according to the present invention is applied to the copying machine. The image processing apparatus includes a main unit 200, a sheet post-processing unit 300 disposed on the left side of the main unit 200, an operation unit 400 for a user to input various operation commands and the like, and an upper part of the main unit 200. The document reading unit 500 is disposed, and the document feeding unit 600 is disposed above the document reading unit 500.

  The operation unit 400 includes a touch panel 401, a start key 402, a numeric keypad 403, and the like. The touch panel 401 displays various operation screens and various operation buttons for the user to input various operation commands. The start key 402 is used for the user to input a print execution command and the like, and the ten key 403 is used for inputting the number of copies to be printed.

  The document feeding unit 600 includes a document placement unit 601, a document discharge unit 602, a paper feed roller 603, a document transport unit 604, and the like, and the document reading unit 500 includes a scanner 501 and the like. The paper feed roller 603 feeds out the original set on the original placement unit 601, and the original transport unit 604 transports the fed originals one by one on the scanner 501. The scanner 501 sequentially reads the conveyed document, and the read document is discharged to the document discharge unit 602.

  The main body 200 includes a plurality of paper feed cassettes 201, a plurality of paper feed rollers 202, a transfer roller 203, an intermediate transfer body roller 204, a photosensitive drum 205, an exposure device 206, and development for each color of yellow, magenta, cyan, and black. Devices 207Y, 207M, 207C, and 207K, a fixing roller 208, a discharge port 209, a discharge tray 210, and the like are provided.

  The photosensitive drum 205 is uniformly charged by a charging device (not shown) while rotating in the direction of the arrow. The exposure device 206 converts the modulation signal generated based on the image data of the original read by the original reading unit 500 into a laser beam and outputs it, and forms an electrostatic latent image on the photosensitive drum 205 for each color. The developing devices 207Y, 207M, 207C, and 207K supply each color developer to the photosensitive drum 205 to form a toner image for each color. Each pixel of the document image data is composed of R (red), G (green), and B (blue) color components, and each color component has a gradation of 0 to 255.

  The intermediate transfer roller 204 transfers the color toner images from the photosensitive drum 205, and forms a color toner image on the intermediate transfer roller 204.

  On the other hand, the paper feed roller 202 pulls out the recording paper from the paper feed cassette 201 in which the recording paper is stored, and feeds the recording paper to the transfer roller 203. The transfer roller 203 transfers the toner image on the intermediate transfer body roller 204 onto the conveyed recording paper, and the fixing roller 208 heats and fixes the transferred toner image on the recording paper. Thereafter, the recording paper is carried into the paper post-processing section 300 from the discharge port 209 of the main body section 200. The recording paper is also discharged to the discharge tray 210 as necessary.

  The paper post-processing unit 300 includes a carry-in port 301, a recording paper transport unit 302, a carry-out port 303, a stack tray 304, and the like. The recording paper transport unit 302 sequentially transports the recording paper carried into the carry-in port 301 from the discharge port 209 and finally ejects the recording paper from the carry-out port 303 to the stack tray 304. The stack tray 304 is configured to move up and down in the direction of the arrow in accordance with the number of recording sheets stacked from the carry-out port 303.

  FIG. 2 is a block diagram showing an electrical configuration of the copying machine shown in FIG. The image forming unit 30 includes a transfer roller 203, an intermediate transfer member roller 204, a photosensitive drum 205, an exposure device 206, and developing devices 207Y, 207M, 207C, and 207K for yellow, magenta, cyan, and black colors illustrated in FIG. The fixing roller 208 corresponds.

  The control unit 10 includes a CPU (Central Processing Unit), a RAM (Random Access Memory), a ROM (Read Only Memory), and a dedicated hardware circuit. The ROM stores a control program for controlling the image processing apparatus, and the CPU executes the control program, whereby the control unit 10 is changed to the character area extraction unit 11, the multi-value conversion unit 12, and the threshold value determination. Unit 13, binarization processing unit 14, output unit 15, and filter processing unit 16.

The character area extraction unit 11 extracts a character area, which is an area composed of pixels constituting a character, from the document image read by the document reading unit 500. Here, the character area extraction unit 11 calculates a difference value of luminance between the target pixel and eight neighboring pixels adjacent to the target pixel, and the difference value having a value larger than a predetermined threshold is equal to or greater than a certain number. If there are (for example, four or more), it is determined that the target pixel is a pixel constituting a character, and a character region is extracted. Here, the luminance is expressed by, for example, Expression (1).
Y = 0.299R + 0.587G + 0.114B Formula (1)
However, Y represents luminance, R represents the gradation of the red component, G represents the gradation of the green component, and B represents the gradation of the blue component. By using the luminance represented by the expression (1), it is possible to extract a character region in consideration of human vision.

  The filter processing unit 16 performs smoothing using, for example, a smoothing filter such as a median filter or an averaging filter on the character region as pre-processing before multi-value conversion of the character region extracted by the character region extraction unit 11. The noise included in the character area is removed. Thereby, the multi-value conversion part 12 can multi-value a character area more correctly.

  The multi-value conversion unit 12 sets the luminance of each pixel to 16 at regular intervals such as 0 to 15 (0 gradation), 16 to 31 (1 gradation),... 239 to 255 (15 gradations). The character image is converted into 16 values by dividing into gradations.

  The threshold value determination unit 13 determines, for each class obtained by dividing 0 to 255 into 16 equal parts, whether or not there are a predetermined number of pixels in a 16-valued character area having gradations belonging to that class. Then, based on the determination result, a threshold for binarizing the document image is determined. FIG. 3 is a table exemplifying the determination result of the threshold value determination unit 13. Each class (gradation) obtained by dividing 0 to 256 into 16 equal parts, and pixels in the character region 16-valued for each class. Presence / absence of the presence and the binarized result are described. In this table, “◯” is described when a predetermined number (for example, 1) of 16-valued character area pixels exist for a certain class, and “X” is described when the number is less than the predetermined number (for example, 0). Yes. For example, the threshold value determination unit 13 sequentially checks whether or not pixels belonging to a certain class exist in the character area, and when there is a predetermined number of pixels, determines that there is a pixel for the class and ends the determination of the class. The determination of “◯” and “×” for each gradation is performed in such a manner that the determination for the next class is performed.

  In the case of Example 1, since “x” is obtained up to the 8th gradation and “◯” is obtained from the 9th gradation to the 15th gradation, the threshold determination unit 13 determines the ninth gradation as the threshold. In the case of Example 2, “◯” appears in the second gradation, but “X” appears again from the third gradation to the fourth gradation, and then “◯” continues from the fifth gradation to the 15th gradation. Therefore, “◯” in the second gradation is determined to be noise, and the fifth gradation is determined as a threshold value.

  The binarization processing unit 14 assigns “1”, assuming that pixels having a gradation equal to or higher than the threshold determined by the threshold determination unit 13 are character parts, and pixels having a gradation lower than the threshold are background portions of the document. “0” is assigned and the character area is binarized. In Example 1 shown in FIG. 3, since the ninth gradation is determined as the threshold value, the binarization processing unit 14 sets “0” to 0 to 8 gradations and “1” to 9 to 15 gradations. Allocation and binarization of character area. In Example 2, since the fifth gradation is determined as the threshold, the binarization processing unit 14 assigns “0” to the 0th to 4th gradations and “1” to the 5th to 15th gradations, The character area is binarized.

  The output unit 15 assigns “0” as the background portion of the document that has not been extracted as the character region by the character region extraction unit 11, and combines the region and the binarized character region to generate a print image. By creating and controlling the image forming unit 30, the print image is printed on the recording paper.

  Next, processing of the image processing apparatus will be described according to the flowchart shown in FIG. First, a document image is read by the image reading unit 500 (step S1), a character region is extracted from the document image by the character region extraction unit 11 (step S2), and the extracted character region is filtered by the filter processing unit 16. Then, it is smoothed and noise is removed (step S3).

  In step S4, the multi-value quantization unit 12 multi-values the character area smoothed in step S3 into 16 gradations of 0 gradation to 15 gradations. In step S5, the threshold value determination unit 13 determines the presence / absence of a 16-valued character area pixel for each of the gradation levels 0 to 15, and determines the threshold value based on the determination result. Is done. In step S6, the binarization processing unit 14 assigns “1” to pixels having a gradation equal to or higher than the threshold determined in step S5 and “0” to pixels less than the threshold. It is binarized.

  In step S7, a print image is generated based on the character region binarized in step S5 and printed on a recording sheet. Here, the output unit 15 displays a preview of the print image on the touch panel 401 before printing the print image on the recording paper, receives a print permission instruction from the user input via the operation unit 400, and then prints the print image. An image may be printed on recording paper.

  Further, as shown in Example 2 of FIG. 3, after “◯” appears in the second gradation, “×” appears continuously, and “◯” appears again from the fifth gradation. When there are a plurality of candidates (in the second example, the second gradation and the fifth gradation), the binarization processing unit 14 binarizes the character area using each candidate as a threshold, and the output unit 15 outputs 2 for each threshold. A thumbnail image of a print image corresponding to each of the digitized character areas may be generated, and a thumbnail image of each print image may be displayed on the touch panel 401 as a preview. Then, the output unit 15 may cause the print image selected by the selection command from the user input via the operation unit 400 to be printed on the recording paper.

  As described above, according to the present image processing apparatus, the character region is temporarily converted into 16 values, and it is determined whether or not a predetermined number of pixels in the 16-valued character region exist for each gradation. Since the threshold value is determined by determining the minimum brightness of the character part included in the character area based on the result, a thin character such as a pencil is compared to the case where the threshold value is determined based on the histogram as in the past. Even if the frequency of is low, the thin character can be detected more accurately. Further, since the threshold value determination unit 13 determines the threshold value for each gradation depending on whether or not there is a pixel, it is not necessary to create a histogram for determining the threshold value as in the conventional device, and simplification. The binarization processing can be performed at high speed by the processed processing.

  In the above embodiment, the multi-value conversion unit 12 converts the character area into 16 values. However, the present invention is not limited to this, and gradations other than 16 (for example, 4, 8, 32, 64...) It may be multi-valued. FIG. 5 is a table exemplifying the determination result of the threshold value determination unit 13 when the multi-value conversion unit 12 converts the character area into four levels of 0 to 4 gradations. In the case of Example 1, since “◯” appears at the first gradation, the threshold determination unit 13 determines the first gradation as the threshold. Then, the binarization processing unit 14 assigns “1” to 1 to 3 gradations which are gradations equal to or higher than the threshold, and assigns “0” to 0 gradation which is less than the threshold, and binarizes the character area. It has become.

  In the case of Example 2, since “◯” appears at the second gradation, the threshold determination unit 13 determines the threshold at the second gradation. Then, the binarization processing unit 14 assigns “1” to the 2nd and 3rd gradations that are the gradations equal to or higher than the threshold, and assigns “0” to the 0th and 1st gradations that are less than the threshold. Is binarized.

  In the above embodiment, the R, G, and B color components of the image data of the document have gradations of 0 to 255, that is, are represented by an 8-bit bit string. However, the present invention is not limited to this. Instead, the one represented by another bit string such as 10, 16 bits may be adopted.

  Furthermore, although not explicitly shown in the above-described embodiment, the output unit 15 shades the binarized background portion of the thin character, or “memo” or “memo” or the like between the thin character lines. It is also possible to create a print image by displaying a comment indicating that it has been binarized, and to indicate to the user the binarized thin character portion.

  In the above embodiment, the filter processing unit 16 performs the smoothing process on the character region using the smoothing filter, but instead performs the edge extraction process using an edge extraction filter such as a differential filter. You may go. In this case, the difference in gradation between the character portion included in the character region and the portion other than the character appears remarkably, and the multi-value conversion unit 12 can convert the character region to m-value more accurately.

FIG. 2 is a schematic side view mainly showing a mechanical configuration of the copying machine when the image processing apparatus according to the present invention is applied to the copying machine. FIG. 2 is a block diagram showing an electrical configuration of the copying machine shown in FIG. 1. It is the table which illustrated the determination result by the threshold value determination part. It is the flowchart which showed the process of this image processing apparatus. It is the table which illustrated the determination result of the threshold value determination part with respect to the quaternarized character area.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 Control part 11 Character area extraction part 12 Multi-value conversion part 13 Threshold value determination part 14 Value conversion process part 15 Output part 16 Filter process part 30 Image formation part 500 Original reading part

Claims (3)

An image processing apparatus for binarizing a document image represented by n (n ≧ 4 integer) gradations,
A character area extracting means for extracting a character area composed of pixels representing characters from the original image based on the luminance of each pixel;
Multi-value conversion means for converting the extracted character area into an m-value image having a gradation number of m (m ≧ 3 and m <n);
Threshold determination means for determining whether or not a predetermined number of each pixel of the m-valued image exists for each gradation and determining a threshold for binarizing the character area based on the determination result When,
An image processing apparatus comprising: binarizing means for binarizing the character area based on the determined threshold value. Further comprising a smoothing means for performing a smoothing process on the character region using a smoothing filter,
The image processing apparatus according to claim 1, wherein the multi-value conversion unit converts the character region smoothed by the smoothing unit into an m-valued image. The character area further comprises edge extraction means for performing an edge extraction process using an edge extraction filter,
The image processing apparatus according to claim 1, wherein the multi-value conversion unit converts the character region extracted by the edge extraction unit into an m-valued image.

Images