JPH1146300A - Image-processing unit - Google Patents

Abstract

PROBLEM TO BE SOLVED: To attain optimum image processing which is suitable for a character or a pattern on a black background by discriminating whether a thin line object is a character or a pattern on a black background and updating a state variable, based on the discrimination result so as to distinguish a character from a pattern on the black background. SOLUTION: A pattern of an oblique component is extracted from both the horizontal and vertical components, and in the case that both sides of a black pattern are inbetween white patterns, it is extracted as a thin line object. Pattern matching is processed, while increasing a subscanning direction counter and a main scanning direction counter, and a current state variable is compared with that of one preceding line. The comparison processing is sequentially conducted, and the state variable is set in response to the discrimination result of an output PM 1 from a pattern matching section A43 to detect a very thin line on white background, thereby avoiding mis-detection of a mesh line on the pattern. Furthermore, since the state variable is set again through matching of thin line patterns, even a lump of characters is extracted satisfactorily.

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 more particularly, to extracting an edge of a line drawing and using the extraction result to distinguish a character and a picture on a black background and perform an appropriate image processing. The present invention relates to an image processing apparatus that performs the following.

[0002]

2. Description of the Related Art In recent years, image processing apparatuses have tended to have higher functions and higher performance. For example, in order to perform high-precision image processing of input image data, an apparatus is provided which determines whether the content of the image data is a character or a photograph.

For example, according to an apparatus disclosed in Japanese Patent Application Laid-Open No. Hei 3-638787, "image reading apparatus", it is determined whether the content of image data is text or a photograph.
By differentiating the contents of image processing using a plurality of color determination results, faithful image reproduction processing can be performed.

Further, according to the method disclosed in Japanese Patent Application Laid-Open No. 1-137378, "Image processing method", a characteristic amount of an input image signal, for example, a density gradient (edge amount) is extracted.
By processing the filter using the feature amount and the threshold used in dither (gradation processing) differently, the edge is sharply reproduced in the line image area such as the character, and the halftone area such as the photograph is reproduced. In this technology, it is possible to smoothly reproduce binary halftones with two values by suppressing noise and the like.

[0005]

However, according to the above-mentioned prior art, although the reproducibility of each image can be improved, Japanese Patent Laid-Open Publication No. Hei 3-63887 merely changes the threshold value of the unit Laplacian data. In JP-A-1-137378, only the processing is changed in accordance with the color (chromatic / achromatic) level.
It was not possible to distinguish characters on black ground, characters on halftone dots, and even pictures. In other words, there is no one that distinguishes a character on a black background, a character on a halftone dot, a picture, and the like, and performs optimal image processing according to each of them.

SUMMARY OF THE INVENTION The present invention has been made in view of the above, and an object of the present invention is to distinguish characters, patterns, and the like on a black background and to perform optimal image processing according to each of them.

[0007]

According to a first aspect of the present invention, there is provided an image processing apparatus for converting an analog image data into an n-value (n is a natural number of 2 or more) digital data. Conversion means, black pixel area determination means for determining an area formed by a plurality of black pixels around a predetermined black pixel in the digital data as a black pixel area, and black pixel area determination means. First state variable setting means for setting a state variable for the corresponding black pixel when it is determined that the pixel is a black pixel area; a corrected black pixel obtained by extracting and correcting a black pixel from the digital data; A thin line candidate extracting unit for extracting a pattern as a thin line candidate using the corrected white pixel extracted and corrected, and the state variable and the thin line candidate extracted by the thin line candidate extracting unit. With a line candidate to determine which one of the character-pattern on black, in which and a determination and update means for updating the state variable based on the determination result.

According to a second aspect of the present invention, there is provided an image processing apparatus, comprising: n-value conversion means for converting analog image data into n-valued (n is a natural number of 2 or more) digital data; And a black pixel area determining means for determining a region constituted by a plurality of black pixels around the black pixel area as a black pixel area. First state variable setting means for setting a state variable in a pixel, a corrected black pixel extracted and corrected by extracting a black pixel from the digital data, and a corrected white pixel extracted and corrected by extracting a white pixel from the digital data. Using the thin line candidate extraction means for extracting a pattern as a thin line candidate, and the state variable and the thin line candidate extracted by the thin line candidate extraction means, With determining which one of the handle, and judgment based on the judgment result to update the state variable
Update means.

According to a third aspect of the present invention, there is provided the image processing apparatus according to the first or second aspect, further comprising:
The apparatus further includes a second state variable setting means for setting the state variable when a black character on a white background is detected.

According to a fourth aspect of the present invention, there is provided the image processing apparatus according to the first, second, or third aspect, and the thin line candidate extracting means is provided in the first, second, or third aspect. However, based on the state variables, pattern extraction is performed by changing the thin line candidate extraction conditions.

According to a fifth aspect of the present invention, in the image processing apparatus according to the first, second, third, or fourth aspect, the scanning direction of the image is performed in a fixed direction, and the reflection direction of the state variable is reflected. Is P with respect to the processing pixel P (i, J).
(I, j + 1), P (i + 1, J), P (i + 1, J-
1).

According to a sixth aspect of the present invention, in the image processing apparatus of the first, second, third, fourth, or fifth aspect, the thin line candidate extracting means uses the Laplacian data to generate the black pixel. And correction of white pixels.

According to a seventh aspect of the present invention, there is provided the image processing apparatus according to the sixth aspect, wherein the thin line candidate extracting means extracts the Laplacian data as the Laplacian data by scanning speed, magnification, or image rotation. Are different.

According to an eighth aspect of the present invention, in the image processing apparatus according to any one of the first to seventh aspects, a character area in the digital data is further determined to determine a character area in the digital data. First character area determining means for outputting a rough state determination result, edge extracting means for extracting an edge in the digital data with a pixel density, and a first character area determining means for determining a character area; And a second character area determining means for determining a true character area when the edge is determined to be an edge by the edge extracting means.

According to a ninth aspect of the present invention, in the image processing apparatus according to any one of the first to seventh aspects, a character area on a halftone dot in the digital data is further determined. A third character area determining means for outputting a determination result in a state rougher than the pixel density; an edge extracting means for extracting an edge in the digital data with the pixel density; And a fourth character area determining means for determining a character area on a true halftone dot when the edge extracting means determines that the character area is an edge.

[0016]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, an image processing apparatus according to the present invention will be described in detail in the order of [Embodiment 1] and [Embodiment 2] with reference to the drawings.

[Embodiment 1] FIG. 1 is a schematic block diagram of an image processing apparatus according to Embodiment 1, which reads image data from a document, converts the image data (analog signal) into digital data, and outputs the digital data. A document reading unit 1, an image processing unit 2 that performs various correction processes and the like on image data (digital data) read by the document reading unit 1, and performs document recognition such as line drawing recognition and color determination; Recording unit 3 for recording an image on recording paper based on image data from
And

Here, the original reading section 1 reads three color image data of R (red), G (green), and B (blue) (hereinafter, referred to as RGB data) and reads the image data. 2 to convert RGB data to C (cyan) M
Color conversion into four color image data (magenta), (yellow), and Bk (black) (hereinafter referred to as CMYBk data), and the image recording unit 3 outputs a color image on recording paper based on the CMYBk data It will be described as what is done.

FIG. 2 is a block diagram of the image processing unit 2, which reads RGB data (digital data) from the original reading unit 1.
To correct the gray balance of the RGB data and to convert the luminance data (RGB data) to the density data (R
RGBγ) which is converted to RGB data and output to the memory 12.
Correction unit 11 and RG of half size (A4) of A3 original
A memory 12 for storing B data;
Based on the RGB data input from the memory 12 and the delay unit 13 for inputting the B data and delaying the RGB data in order to synchronize with the output result of the document recognizing unit 14, the character area or the picture area is determined. Next-stage RGB filter section 15
And a character / picture determination signal C / P, and outputs a chromatic / achromatic determination signal B / C to the RGB filter section 15 by determining whether the document area is a chromatic area or an achromatic area. 14, an RGB filter unit 15 for performing MTF correction on the RGB data, and CM conversion of the RGB data by primary masking or the like.
A color correction unit 16 for converting into Y data, a UCR unit 17 for generating a Bk data by performing a UCR (additive color removal) process on a common part of the CMY data, and a conversion unit for performing enlargement / reduction or equal magnification processing in the main scanning direction. A multiplying unit 18, a CMYBk filter unit 19 for performing a smoothing process and a sharpening process, and a CMYBkγ correcting unit 20 for performing a γ correction according to the frequency characteristics of the image recording unit 3.
And a gradation processing unit 21 for performing quantization such as dither processing and error diffusion processing.

The character / picture determination signal C / P output from the document recognizing section 14 is a 2-bit signal.
When the pattern determination signal C / P is “3” = “11”, the character area is indicated, “1” = “01” indicates a character on the pattern, and “0” = “0”.
"0" indicates a picture area. This character / picture determination signal C / P
Are the RGB filter unit 15, the color correction unit 16, and the UCR unit 1.
7, scaling unit 18, CMYBk filter unit 19, CMYB
It is cascaded to the kγ correction unit 20 and the gradation processing unit 21 and outputs a signal IMG in synchronization with image data.

In the logic of the chromatic / achromatic determination signal B / C (1 bit signal), H indicates an achromatic region and L indicates a chromatic region. The chromatic / achromatic determination signal B / C is cascaded to the RGB filter unit 15, the color correction unit 16, and the UCR unit 17, and is output in synchronization with image data.

Each of the blocks 15 to 21 performs character processing and picture processing based on the character / picture determination signal C / P and the chromatic / achromatic determination signal B / C.

Next, the detailed operation of each block of the image processing section 2 will be described.

The RGBγ correction section 11 corrects the gray balance of the RGB data (digital data) input from the document reading section 1 and converts the luminance data (RGB data) into density data (RGB data). 1
Output to 2.

The memory 12 is a two-chamfered memory.
Convert image data (RGB data) to maximum document size (A3)
Of the image (A4), and the rotation of the image is also performed using this memory. Note that the memory 12 is arranged near the document reading unit 1 so that all images can be formed by reading the document once. In addition, by arranging the memory 12 and the document reading unit 1 close to each other, it is possible to output data with less reading variation to the memory 12. The control of the memory 12 will be described later.

The RGB filter section 15 is an N × N pixel filter for MTF-correcting the RGB data from the delay section 13, and performs a sharpening process when the character / picture determination signal C / P is “3” (character). When the value is "0" (pattern), the smoothing process is performed. When the value is "1" (character on the pattern), the input data is output without processing.

The color correction section 16 converts the RGB data processed by the RGB filter section 15 into C, M, Y data by a primary masking method or the like.

The UCR unit 17 converts the C, C converted by the color correction unit 16 to improve the color reproducibility of the image data.
Bk data is generated by subjecting a common portion of the M and Y data to UCR (additive color removal) processing. Specifically, UCR
The unit 17 generates skeleton black Bk data when the character / picture determination signal C / P is other than “3”, and generates “3”
, Full black Bk data is generated, and when the character / picture determination signal C / P is “3” (character) and the chromatic / achromatic determination signal B / C is H (achromatic), black characters are generated. The C, M, and Y data are erased to represent only the black component. Further, the UCR section 17 sends the image reading section 1 four times to the scaling section 18 to read one original document four times.
To output these in C-M,
A signal IMG of one color of Y and Bk is output in a frame sequence.

The magnification unit 18 enlarges / reduces in the main scanning direction,
Perform 1: 1 processing. The CMYBk filter unit 19 is a filter of N × N pixels and performs a sharpening process or a smoothing process based on the frequency characteristics of the image recording unit 3 and the character / picture determination signal C / P. The CMYBkγ correction unit 20 is the image recording unit 3
Γ based on the frequency characteristics of the
The curve is changed, and specifically, the character / pattern determination signal C / P
Is “0” (picture), γ that faithfully reproduces the image
When the value is other than "0", the contrast is enhanced by setting γ.

The gradation processing section 21 performs quantization such as dither processing based on the gradation characteristics of the image recording section 3 and the character / picture determination signal C / P. When the value is “0” (picture), gradation-oriented processing is performed and “0”
At times other than the above, resolution-oriented processing is performed.

A summary of the character / picture processing of each of the blocks 15 to 21 based on the character / picture decision signal C / P and the chromatic / achromatic decision signal B / C is shown below.

When the character / picture determination signal C / P = “0” (picture) RGB filter section 15 → smoothing process UCR section 17 → skeleton black CMYBk filter section 19 → γ curve emphasizing linear (gradation) Selection CMYBk γ correction unit 20 → Gamma curve selection for faithfully reproducing images Gradation processing unit 21 → Gradation emphasis

When character / picture determination signal C / P = “3” (character) RGB filter section 15 → enhancement processing UCR section 17 → full black CMYBk filter section 19 → enhancement processing CMYBkγ correction section 20 → contrast by setting γ Tone processing unit 21 → emphasized resolution

When character / pattern determination signal C / P = “3” (character) and chromatic / achromatic determination signal B / C = H (black character processing) Prevention of color shift due to misalignment around black characters Do not print YMC data. In this case, B
The k-data RGB filter may be performed more strongly than in the case of color characters to make it clearer.

When the character / picture determination signal C / P is "1" (character on the picture) RGB filter section 15 → weak emphasis processing or through processing for directly outputting input data UCR section 17 → full black, in this case, YMC data is printed without processing such as black character processing. CMYBk filter section 19 → enhancement processing CMYBkγ correction section 20 → contrast enhancement by setting γ gradation processing section 21 → resolution emphasis

As described above, the image processing section 2 can perform three types of processing of a pattern, a character, and a character on a pattern.

FIG. 3 shows a document recognition unit 1 which is a main part of the present invention.
FIG.

As shown in FIG. 3, the document recognizing section 14 is roughly divided into a line drawing recognizing section 14a for detecting the likeness of a line drawing and a color determining section for determining whether a specific area of the document is chromatic or achromatic. 14b. Here, a case where the reading density of the document reading unit 1 is about 400 dpi will be described as an example.

As shown, the line drawing recognition unit 14a includes a monochrome conversion unit 41, a Laplacian unit 42, pattern matching units A43, B44, C45, and an isolated point removal unit A4.
6, B47, pixel density converters A48, B49, page memories A50, B51, selectors A52, B53.
, An isolated block removing section A54, B55, and an expanding section A5
6, B57, AND operators 58 and 59, and an edge extraction unit 14c. This edge extraction unit 14c
Is a white isolated point removing unit 61, a black isolated point removing unit 62, a contour extracting unit 63, an expansion / reduction unit 64, and a line delay 6
Consists of five.

In the line drawing recognition section 14a, the expansion section A5
6 and the edge extraction unit 14c are output from the system A as the MSB (C / PA) of the character / picture determination signal C / P, while the AND output of the expansion unit B57 and the edge extraction unit 14c is Character / Picture determination signal C / P from system B
Is output as the LSB (C / PB). That is, the line drawing recognition unit 13a outputs C / PA and C / PB,
When PA, C / PB is (L, L), the character / design determination signal C / P is “0”, and when PA, C / PB is (L, H), “1”, (H, At the time of H), the character / picture determination signal C / P is set to "3", and C / PA and C / PB are called the character / picture determination signal C / P.

The color judgment section 14b outputs a chromatic / achromatic judgment signal B
/ C is output, and its output logic outputs L when it is a chromatic area and outputs H when it is an achromatic area. The output result is a signal in which 4 × 4 pixels correspond to one pixel.
Hereinafter, the unit of the output result is one block. The color determination unit 14b includes a color determination circuit 66 as shown in the figure.
, A page memory C67, and a selector C68.

In the above configuration, the operation of each part of the line drawing recognition unit 14a, which is a main part of the present invention, will be described in detail.

The monochrome conversion section 41 converts the RGB data into luminance data or the like, converts the data into monochrome, and outputs the monochrome data to the Laplacian section 42. Note that instead of the luminance data,
The darkest data among the RGB data may be selected, or the G data may be used as the luminance data.
Here, the output data of the monochrome conversion unit 41 indicates that the larger the numerical value is, the deeper the data is, and the smaller the numerical value is, the lighter the output data is.

The Laplacian section 42 extracts the edge of the line drawing, detects a white region and a black region, and detects a white (w) region to obtain extracted data of a line drawing on a white background. Is detected, and is used as extracted data of a line drawing on a black character.

Next, the white area extraction operation and the black area extraction operation by the Laplacian unit 42 will be described with reference to FIGS. FIG. 4 is an explanatory diagram showing reference pixels of the document recognition unit in FIG. 3; FIG. 5 is a cross-sectional view of a line drawing (an explanatory diagram showing a density change of image data), and is a conceptual diagram showing a relationship between a white region, a black region, and a threshold. In FIG. 5, THb is the threshold value of the black area,
thw1 indicates a threshold for a thin line in a white area, thw2 indicates a threshold for a thick line in a white area, and thw3 indicates a threshold for correction of a white area. FIG. 6 is an explanatory diagram showing the relationship between the edge amounts (x) of the image data in FIG.

First, the operation of white area detection will be described. For example, a 3 × 3 pixel block in the matrix shown in FIG.
When the peripheral data including 11 is smaller than the threshold thw, it is determined as a white area candidate. ｛(A00 <thw) and (a01 <thw) and (a02 <thw
) And (a10 <thw) and (a11 <thw) and (a12
<Thw)｝ or｝ (a10 <thw) and (a11 <thw) and
(A12 <thw) and (a20 <thw) and (a21 <thw
) And (a22 <thw)｝ or ｛(a00 <thw) and (a
10 <thw) and (a20 <thw) and (a01 <thw) and
(A11 <thw) and (a21 <thw) {or} (a01 <thw
) And (a11 <thw) and (a21 <thw) and (a02)
<Thw) and (a12 <thw) and (a22 <thw)｝

Here, the threshold value thw for detecting the white area candidate for the thick line and the thin line uses a different value, and as shown in FIG. 5, the general white background value thw2 is used for the "white area candidate for the thick line".
Is set, and in the “thin line white area candidate”, a value thw1 that is slightly lower (closer to white) than a general white background is set. The reason is to avoid the highlight (brighter one) of the picture (halftone dot of the printed matter or line of the copying machine) from becoming a “white area candidate”. Although this pattern shows an example of an orthogonal pattern, an oblique pattern may be added.

Next, in order to further calculate a white area from the white area candidates, a Laplacian is obtained as in the following equation (1). x1 = a22 * 2- (a21 + a23) * ix2 = {a22 * 4- (a11 + a13 + a31 + a33)} * i / 2 + x1 x3 = a22 * 2- (a12 + a32) + x2 (1)

Here, i is M in the main scanning direction and the sub-scanning direction.
It is a weighting factor for performing a difference between TFs and correction during zooming and rotation. If the value of x at this time is as follows, it is regarded as a white area. x−N <x3 <x + N Here, the thresholds for the thick line and the thin line do not have to be separate.

In this manner, the white area for thin lines and the white area for thick lines are detected. As a result, small halftone dots and line patterns on the highlight side of the picture are removed so as not to be extracted.

Next, the correction of the "white area for thick line" will be described. For example, in the case of an original of black-and-white inverted characters (white characters and black surroundings), when an original is read by an optical reading device such as a copying machine, flare (not one white point but the surroundings are affected by black) ), The white data may be slightly blacker than usual. Therefore, the following correction is performed.

For example, when the matrix of the white area is 3 × 3, the following correction is performed. ｛(A00 <thw) and (a01 <thw) and (a02 <thw
) And (a10 <thw) and (a11 <thw) and (a12 <
thw) and (a20 <thw) and (a21 <thw) and (22
<Thw)｝

This is set as a “correction white area candidate”, and a “correction white area” is calculated from the Laplacian described above. Here, the threshold value thw is thw3 closer to black than the threshold value thw2 of the thick line as shown in FIG. 5, and N is set to be smaller than the value N of the white area for the thick line described above. The reason for reducing N is to extract stable data with a small amount of change in white data. The extracted “corrected white area” is corrected to the above-described “white area for thick line”, and is set as a “corrected white area for thick line”. That is, the “corrected white area for thick line” is the “corrected white area” or the “white area for thick line”. Also in this case, small halftone dots and line patterns on the highlight side on the picture are removed so as not to be extracted.

Next, the black area detecting operation will be described.
A 3 × 3 pixel block in the matrix shown in FIG. 4 is used. First, when the peripheral data including the pixel of interest a11 is larger than the threshold value THb according to the following logic, it is determined as a black area candidate. ｛(A00> THb) and (a01> THb) and (a02> THb)
) And (a10> THb) and (a11> THb) and (a12>
THb)｝ or ｛(a10> THb) and (a11> THb) and
(A12> THb) and (a20> THb) and (a21> THb) a
nd (a22> THb) {or} (a00> THb) and (a10> T)
Hb) and (a20> THb) and (a01> THb) and (a11
> THb) and (a21> THb)｝ or ｛(a01> THb) and
(A11> THb) and (a21> THb) and (a02> THb) a
nd (a12> THb) and (a22> THb)｝

Here, the threshold value THb of the “black area candidate” is shown in FIG.
As shown in (1), a general black background value (in other words, the density to be emphasized as a character) is set (THb> thw3). Although this pattern is an orthogonal pattern, a pattern such as an oblique pattern may be added.

Next, similarly to the case of white, a Laplacian is further obtained from the black region candidate by the above-described equation (1) in order to calculate a black region.

Further, the edge amount x as shown in FIG. 6 is obtained by the above equation (1). Here, the weighting coefficient i in the equation (1) is a selective coefficient, and a coefficient i = 1 = 1.115 = 1. 25 = 1.375 = 1.5 = 1.625 = 1.775 = 1.875 = 2 (fixed decimal point calculation).

By setting the coefficient i in this manner, blurring of the MTF (optical system and traveling system) in the main scanning direction and the sub-scanning direction can be corrected. Here, the MTF in the main scanning direction and the MTF in the sub-scanning direction are different, and in the copying machine, the magnification in the sub-scanning direction is determined by the reading area (speed) of the reading device.
Varies depending on the magnification. In the memory unit 12, for the 90-degree rotation of the image data, i is also set to a coefficient corresponding to the rotation.

Therefore, in the present embodiment, no problem occurs because the scaling section 17 in the main scanning direction is located after the document recognition section 13. Further, in this embodiment, when the magnification in the sub-scanning direction is large, for example, 200%, the edge amount x may be obtained by making the matrix shown in Expression (1) selectable.

That is, the Laplacian section 42 sends the white area symbol w (for the thick line and the thin line) (H for the white area) and the black area to the pattern matching sections A43, B44, C45 and the edge extraction section 14c at the subsequent stage. Area symbol k (H in black area)
And a signal RAP including the edge amount x.

Next, edge extraction by the edge extraction unit 14c will be described. The edge extraction unit 14c extracts edges from the signal RAP output from the Laplacian unit 42. As shown in FIG. 6, the black pattern (k) is H when the black area or the edge amount x is larger than the threshold value THRb, and the white pattern (w) is other than the black pattern.

First, the white isolated point removing section 61 will be described. The white isolated point removing unit 61 removes isolated white isolated points. Specifically, when the target pixel is white and the periphery is black, the target pixel (white) is corrected to black by the following equation. In the following equation, the central pixel is w22. (K00 and k01 and k02 and k03 and k10 and k13 and k20 and w22 and k23 and k30 and k31 and k32 and k33) or (k01 and k02 and k03 and k04 and k11 and k14 and k21 and w22 and k24 and k31 and k32 and k33 and k34) or (k10 and k11 and k12 and k13 and k20 and w22 and k23 and k30 and k33 and k40 and k41 and k42 and k43) or (k11 and k12 and k13 and k14 and k21 and w22 and k24 and k31 and k34 and k41 and k42 and k43 and k44)

Next, the black isolated point removing section 62 will be described. The black isolated point removing unit 62 removes isolated black isolated points. Specifically, when the target pixel is black and the periphery is white, the target pixel (black) is corrected to white by the following expression. In the following equation, the pixel of interest is k01. (K00 and k01 and k02 and k10 and k11 and k12) or (k00 and k01 and k10 and k11 and k20 and k21)

Next, the contour extracting unit 63 will be described. The contour extraction unit 63 performs contour extraction. Since the black and white isolated points have been removed, a good edge can be extracted. An example of the contour extraction pattern is shown below. w11 and (k00 or k01 or k02 or k10 or k12 or k20
or k21 or k22)

Next, the expansion / reduction unit 64 will be described.
The expansion / reduction unit 64 performs expansion / reduction of the contour. Specifically, the expansion / reduction unit 64 performs an N × N OR process (expansion process) to connect an isolated block to an adjacent or neighboring block, and then performs an M × M AND process ( (Shrinkage processing). Then, for example, 5 × 5 interpolation processing is performed. Note that NM is an expansion process.

As an example, a 7 × 7 (N = 7) expansion process will be described below. a00 or a01 or a02 or a03 or a04 or a05 or a06 or a10 or a11 or a12 or a13 or a14 or a15 or a16 or a20 or a21 or a22 or a23 or a24 or a25 or a26 or a30 or a31 or a32 or a33 or a34 or a35 or a36 or a40 or a41 or a42 or a43 or a44 or a45 or a46

Further, as an example, a contraction process of a 3 × 3 (M = 3) block will be described below. a00 and a01 and a02 and a10 and a11 and a12 and a20 and a21 and a22

Thus, the isolated area can be reduced and expanded.

The line delay section 65 includes an expansion section A56,
In order to synchronize with the output result of B57, the output from the expansion / contraction section is delayed and output to AND operators 58 and 59.

Next, the pattern matching units A43, B4
4 and C45 will be described.

The pattern matching unit A43 extracts a white area around the black area. Where white area pattern (W)
Is the signal of the white area for the corrected thick line, and the black pattern (K)
Is a black area signal. As an example of the pattern,
× 7).

[K12 and k13 and k14 and k22 and k23
and k24and k32 and k33 and k34and ｛(w52 and w5
3 and w54) or (w62 and w63 and w64) or (w12 and
w13 and w14) or (w02 and w03 and w04)｝] or [k
21 and k31 and k41 and k22 and k32 and k42and k23
and k33 and k43and ｛(w25 and w35 and w45) or (
w26 and w36 and w46) or (w21 and w31 and w41) or
(W20 and w30 and w40)｝]

In the above example, only the horizontal component and the vertical component are shown, but similarly, the pattern of the oblique component is extracted. In this way, a white area on a black area is extracted. Since there are many black areas, it is possible to extract a line image of a black area without erroneously recognizing a halftone dot as a line image.

The coding may be performed by utilizing the magnitude relation of the black area, the white area for thick line correction, and the white area of the thin line. As an example of the coding, the black area is B, the white area for thick line correction is W1,
The description will be made assuming that the thin line white area is W2. In this case, if it is not coded, it becomes 3 bits × n lines, but if it is coded as follows, it becomes 2 bits × n lines. When B → Code “1” = (01) When W2 → Code “2” = (02) When W1 and not W2 → Code “3” = (11) When neither B nor W1 or W2 → Code “ 0 ”= (00)

Since the code is "0" to "3", it can be represented by 2 bits. When the code is expanded, the reverse process may be performed. In addition, the magnitude relation need not be fixed, but it is of course better to be able to exchange the magnitude relation.

The processing flow of the putting section A43 is as shown in the flowcharts of FIGS. 10 and 11. As shown in FIGS. 8 and 9, the main scanning direction X (=
As shown in the flowchart of FIG. 10, the pattern matching process is performed while incrementing the sub-scanning direction counter j and the main scanning direction counter i at i) and the target pixel Pij in the sub-scanning direction Y (= j).

Here, the pattern matching process (the pattern matching process in FIG. 10) will be described with reference to the flowchart in FIG. First, pattern matching is performed to determine whether or not the pattern matches the above-mentioned pattern (S1). If the pattern matches, the output PM1 (see FIG. 2) outputs H (on) (S2). off)
Is output (S3).

The pattern matching unit A43 extracts the edge of the thick line portion of the line drawing by the above processing. This output signal PM1 is supplied to the pattern matching units B44 and C4.
5, applied to the isolated point removing units A46 and B47.

The pattern matching section B44 detects a fine line. Here, a thin line means a character or a line drawing having a line width of 1 mm or less. As shown in FIG. 6, the black pattern (k) is H when the black area or the edge x is larger than the threshold THRb, and the white pattern (w) is the white area for thin lines or the edge x is smaller than the threshold THRw. (The absolute value is large because it is a negative component) is set to H. The thresholds THRb and THRw may be changed by the magnification, the type of the original (color, black and white, print photograph, photographic paper photograph, copy original, map, etc.), an adjustment key, or the like. The reason for the correction with the edge component is to increase the contrast of the thin line.

As an example of the fine line pattern, the following (7)
× 7). ｛(W22 and w23 and w24) or (w02 and w03 and w
04) and w12 and w13 and w14and k32 and k33 and k34
and w52 and w53 and w54and (w42 and w43 and w44)
or (w62 and w63 and w64)｝ or ｛(w22 and w32 a
nd w42) or (w20 and w30 and w40) and w21 and w3
1 and w41and k23 and k33 and k43and w25 and w35 an
d w45and (w24 and w34 and w44) or (w26 and w36
and w46)｝ or [(w12 and w13 and w14) or (w02
and w03 and w04) and k32 and k33 and k34and ｛(k
22 and k23 and k24) or (k42 and k43 and k44)｝ an
d (w52 and w53 and w54) or (w62 and w63 and w6
4)] or [(w21 and w31and w41) or (w20 and w3
0 and w40) and k23 and k33 and k43and ｛(k22 and
k32 and k42) or (k24 and k34 and k44)｝ and (w
25 and w35 and w45) or (w26 and w36 and w46)]

Although only the horizontal component and the vertical component are shown here, the pattern of the diagonal component is similarly extracted. As described above, when both sides of the black pattern are sandwiched by the white patterns, they are extracted as thin line candidates.

The processing flow of the pattern matching unit B44 is as shown in the flowcharts of FIGS. As shown in FIGS. 8 and 9, at the pixel of interest Pij in the main scanning direction X (= i) and the sub-scanning direction Y (= j), as shown in the flowchart of FIG. The pattern matching process is performed while incrementing the scanning direction counter i.

Here, the pattern matching process (the pattern matching process in FIG. 10) will be described with reference to the flowchart in FIG. Note that MFB is a state variable, and is 0 at the leading end of main scanning. SFB (i) is an array for one line in the main scanning direction, and is a state variable one line before.

First, the current state variable MFB is compared with the state variable SFB (i) one line before (S11). Here, if MFB <SFB (i), MFB = SFB
(I) (S12), the process proceeds to step S12; otherwise, the process directly proceeds to step S12.

In step S12, it is determined whether or not the output PM1 from the pattern matching unit A43 is on. If PM1 is not on (ie, if it is off), the process proceeds to step S18. On the other hand, if PM1 is on, if the value of MFB is larger than 0 in steps S14 to S17, the value is changed. Specifically, MFB is 8
If it is larger, it is set to 16 (S14, S15), and MF
If B is greater than 0 and less than 8, it is set to 8.

In step S18, it is determined whether or not the area is a white background area. The white background area here is the Laplacian part 4
The thin line white area of the output of No. 2 is assumed to be a, and it is determined to be a white background area when:

A00 and a01 and a02 and a10 and a11 and a12 and a20 and a21 and a11

If it is determined in step S18 that the area is a white background area, L (off) is output to outputs 1 and 2 of the pattern matching unit B44 (S19), and MFB =
The value is set to 16 (S20), and the process proceeds to step S36.

On the other hand, if it is determined in step S18 that the area is not a white background area, it is determined whether or not the area is a thin line pattern by determining whether or not the area matches the above-described thin line pattern (S21). L (off) is applied to output 1 and output 2 of the pattern matching unit B44.
Is output (S22), and it is determined whether or not MFB = 0 (S23). If MFB = 0, the process proceeds to step S36. If not MFB = 0, MFB = MFB-1 is set ( S24), and proceed to step S36.

If it is determined in step S21 that the pattern is a thin line pattern, it is determined whether MFB> 8 (S25). If MFB> 8, the output 1 and output 2 of the pattern matching unit B44 are set to H (on). ) Is output (S
26), and in steps S28 and S29, MFB> 1
If MFB = 6, MFB = 16 is set, and if MFB> 16, the process proceeds directly to step S36.

If it is not larger than 8 in step S25, it is determined whether or not MFB = 0 (S30).
If FB = 0, L (off) is output to output 1 and output 2 of pattern matching section 305B (S31),
Proceeding to step S36, if MFB is not 0, L (off) is output as output 1 of the pattern matching unit 305B, H (on) is output as output 2 (S32), and MFB =
MFB + 4 (however, if MFB is 16 or more, set to 16) (S33), and in steps S34 and S35, if MFB> 8, set MFB = 8 if MFB> 8, and set MFB> 8 if MFB> 8. If not, step S36
Proceed to.

In step S36, SFB (i) = MF
B, and updates the state variable SFB (i) one line before. Next, in step S37, SFB (i)> SFB
(I-1) is determined. This is a comparison between the updated data one line before and the updated data one pixel before one line. If the data SFB (i) one line before is large,
SFB (i-1) = SFB (i) is set (S38),
The process ends.

When this processing is sequentially performed in the main scanning direction, the state variable MFB propagates in three arrow directions shown in FIG. That is, the update process (SFB) shown in step S36
(I) = MFB), P (i,
j) to P (i + 1, j), and by the processing {SFB (i) = SFB (i−1)} shown in step S38,
As shown in the diagonally lower left direction, P (i, j) to P (i +
i, j + 1). From this, step S1
8 or the determination result of the output PM1 from the pattern matching unit A43 in step S13,
By setting the state variables, it becomes possible to detect a fine line on a white background, and it is possible to prevent halftone dots on a picture from being erroneously extracted. Furthermore, since the state variables are reset by the thin line pattern matching in step S21, it is possible to satisfactorily extract chunks of characters.

Since the output and output 2 of the pattern matching unit B44 are output differently using the state variables, it is possible to output the characters on the white background and the characters on the halftone dots separately.

Further, as is apparent from FIG. 9, since the arrow direction of the sub-scan is 0 or + (plus) direction, the processing performed for each line (each main scanning line) requires It is sufficient to provide only the required number of lines of memory for variables and pattern matching, and it can be easily realized without providing a page memory or the like.

As shown in FIG. 2, the pattern matching section B44 outputs the output 1 to the isolated point removing section A46, and outputs the output 2 to the isolated point removing section B47. The difference between output 1 and output 2 is the difference in state variables. Thereby, for example, as shown in FIG. 7, characters (A, I, U, etc.) are described inside the frame of the ruled line, and when the inside of the frame is a halftone dot, the frame of the ruled line is output 1, The output 2 is determined to be a thin line, and only the output 2 having a large state variable can be determined to be a thin line for a character on a halftone dot.

The pattern matching section C45 detects a reversed fine line. Here, the purpose is to detect a thin white line on a black ground. The hardware configuration is the same as that of the pattern matching units A43 and B44. However, the black and white of the input data are reversed. The logic of the thin line pattern and the white background area are reversed, and the white area is a black area. In the pattern matching, a white area surrounded by a black area is obtained, and a threshold value to be extracted is separately required. By doing so, it is possible to satisfactorily extract the inverted fine line. The pattern matching unit C45 detects the inverted fine line and outputs it to the isolated point removing units A46 and B47.

Next, the isolated point removing units A46 and B47 will be described. Both the isolated point removing units A46 and B47 are composed of the same circuit. The input data of the isolated point removing unit A46 is the output PM of the pattern matching unit A43, the output 1 of the pattern matching unit B44, and the output of the pattern matching unit C45. The input data of the isolated point removing unit B47 is the pattern matching unit A43. Output PM1,
These are the output 2 of the pattern matching unit B44 and the output of the pattern matching unit C45. Isolated point removing unit A46,
In B47, since the line drawing is composed of continuous lines, isolated points are removed. An isolated point occurs when a halftone dot is erroneously detected as a line image.

In the isolated point removing units A46 and B47,
If any one of the pattern matching units A43, B44, C45 is an extraction pattern, it is determined as an extraction pattern. For example, when the number of extracted patterns is two or more in pattern matching using a 4 × 4 matrix, the isolated point removing units A46 and B47 correct the center pixel (a22 or a33) as an extracted pattern and output H. Is output (as an extraction pattern). As a result, the isolated points are removed and, at the same time, they are expanded (expanded). As shown in FIG. 3, the outputs of the isolated point removing units A46 and B47 are PM2,
PM3.

Next, the pixel density conversion units A48 and B49 will be described. The pixel density conversion units A48 and B49 are also formed of the same logic circuit. Here, in the circuits up to the previous stage, processing is performed in pixel units. However, in the pixel density conversion units A48 and B49, in order to perform processing in block (4 × 4 pixel) units, data in pixel units is converted into data in block units. Convert. In this conversion, a simple 4 × 4 thinning-out process is performed. However, since the isolated point removing units A48 and B49 also perform a substantially 4 × 4 expansion process, data loss does not occur. The output of the pixel density converter A48 is applied to the page memory A50 and the selector A52, and the output of the pixel density converter B49 is applied to the page memory B51 and the selector B53.
Is applied to

Next, the page memories A50, B51, and C67 (page memory of the color determination unit 14b) will be described. The circuits of the page memories A50, B51 and C67 all have the same function. The page memory A50 is
The output result of the pixel density conversion unit A48 is input, and the page memory B51 receives the output result of the pixel density conversion unit B49.
The page memory C67 receives the output result of the color determination circuit 66.

The page memories A50, B51 and C67
Is, for example, 1200 dots in the main scanning direction × 1 in the sub-scanning direction.
It is composed of 736 lines (approximately 2 MB), and has an A3 size with a resolution of 16 dots / mm in both the main and sub-scanning directions and a size larger than DLT paper. At the time of the first scan, the input data is stored in the page memories A50, B51 and C67 in block (4 × 4 pixel) units, and at the same time, is output via the selectors A52, B53 and C68. After the second scan, the judgment results stored in the page memories A50, B51 and C67 at the time of the first scan are output via the selectors A52, B53 and C67. That is, since the color determination result in the first scan and the line drawing extraction processing data are used in the second and subsequent scans, there is no variation in the color determination result (B / C) and line drawing extraction result (C / P) for each scan. Can be.

Next, the isolated point removing units A54 and B55 will be described. The isolated block removing units A54 and B55 are the same circuit and have the same function. Selector A52,
In the output data from B53, blocks isolated from the peripheral data are removed. Specifically, for example, 5
In a matrix of × 5 blocks, when only the center block is on (H) and the others are off (L), this block is isolated, and the output L is output as off. o
n means an extraction pattern. As a result, blocks isolated from the peripheral data are removed.

Next, the expansion portions A58 and B59 will be described. The expansion portions A58 and B59 have the same function in the same circuit. Here, N × N OR processing (expansion) is performed, and then M × M AND processing is performed (reduction). And 5
× 5 interpolation processing is performed. MN is the expansion amount.

M and N satisfy N> M. The reason why the OR processing is performed here is to connect an isolated block to an adjacent or neighboring block. As an example, an example of expansion of a 3 × 3 block (corresponding to 12 × 12 pixels) is shown. a00 or a01 or a02 or a10 or a11 or a12 or a20 or a21 or a22

Thereafter, AND processing (reduction) of 5 × 5 pixels is performed. An example is shown below. a00 and a01 and a02 and a03 and a04 and a10 and a11 and a12 and a13 and a14 and a20 and a21 and a22 and a23 and a24 and a30 and a31 and a32 and a33 and a34 and a40 and a41 and a42 and a43 and a44

Thereafter, since the jaggies of 100 dpi remain, interpolation processing is performed. FIG. 13 shows an example of the interpolation processing. In FIG. 13, the solid line indicates the jaggedness before the correction of 100 dpi, and the broken line indicates the output after the correction. For example, 5 ×
In a matrix of 5: When the line drawing extracted data satisfies the following logical expression, the data of the target pixel a22 is reflected (corrected). (Pattern 1 and! Pattern 2) or (Pattern 3 and!
Pattern 4) Details of patterns 1, 2, 3, and 4 are as follows. In addition, here! Indicates an indefinite operator.

Pattern 1 (a00 and a02 and! A04 and a20 and a22 and! A24a
nd! a40 and! a42 and! a44) or (! a00 and a02a
nd a04 and! a20 and a22 and a24and! a40 and! a4
2 and! a44) or (! a00 and! a02 and! a04 and a2
0 and a22 and! a24and a40 and a42 and! a44) or
(! A00 and! A02 and! A04 and! A20 and a22 and
a24and! a40 and a42 and a44)

Pattern 2 (a11 and a12 and a13 and a21 and a22 and a23and a
31 and a32 and a33)

Pattern 3 (! A00 and! A02 and a04 and! A20 and! A22 and
a24and a40 and a42 and a44) or (a00 and! a02
and! a04 and a20 and! a22 and! a24and a40 and
a42 and a44) or (a00 and a02 and a04 and! a2
0 and! a22 and a24and! a40 and! a42 and a44) o
r (a00vand a02 and a04 and a20 and! a22 and! a24
and a40 and! a42 and! a44)

Pattern 4 (! A11 and! A12 and! A13 and! A21 and! A22 an
d! a23and! a31 and! a32 and! a33)

As described above, by expanding the extracted pattern, the intersections of the characters and the like are connected, and the line drawing and the periphery thereof are subjected to line drawing processing. In the above-described pattern matching, cross points of crosses cannot be extracted, but they can be connected by this dilation processing. The line drawing and its surroundings are regarded as line drawing because
This is to make the black character processing and the spatial filter work well.

As a result, a ruled line can be satisfactorily extracted even if there is a halftone dot in the ruled line as in the catalog usage period / specification table. In addition, the ruled lines including the white background are extracted as characters, and the characters on the halftone dot are output as different judgment results from the characters on the white background. It is possible to do. Furthermore, the white thin line on the black character is also good.

The AND operators 58 and 59 respectively output the outputs of the expansion units A56 and B57 and the edge extraction unit 14c.
Is output from the system A as the MSB (C / PA) of the character / pattern determination signal C / P.
LSB (C / P) of character / picture determination signal C / P from system B
-Output as B). That is, the character area (C / PA,
B) and the edge of the extracted 400 DPI signal A
By performing the ND, it is possible to satisfactorily extract the edge of the character. Edge position accuracy is improved by extracting edges at 400 DPI. Here, the signal of 400 DPI is not stored in the memory because it is not output from the isolated point removing unit A46 or the isolated point removing unit B47, but is simply stored as an edge. Edge signal.

[Embodiment 2] In Embodiment 2, the output condition of the pattern matching unit B44 of the line drawing recognition unit 14a shown in FIG. 3 of Embodiment 1 is changed. The basic configuration and operation are the same as those in the first embodiment, and only different portions will be described here.

In the first embodiment, in the fine line pattern matching performed by the pattern matching unit B44, the output 1 and the output 2 are set according to the state variable. In the second embodiment, the setting condition of the output 2 is further added. To extract characters (ruled lines) on the picture.

FIG. 14 shows a flowchart of the pattern matching processing (pattern matching section B44) according to the second embodiment. Since the same reference numerals as those in the flowchart of the first embodiment shown in FIG. 10 indicate common processing, only different points will be described here. Steps S11 to S shown in FIG.
38 is the same as FIG. 12, except that steps S21 and S22
Steps S41 to S44 are added between steps S24 and S32.

In step S21, first, it is determined whether or not the pattern is the above-described fine line pattern. If the pattern is not a fine line pattern, it is determined whether or not the pattern matches the following fine line pattern 1 (S41).

As the fine line pattern 1, the same fine line pattern as described above is used, but it need not be the same. Here, the black pattern (k) has a threshold value THRb for the black area or edge amount.
(See FIG. 6) A larger one is designated as H. The white pattern (w) has a thin line white area or an edge amount of a threshold THRw.
Less than (absolute value is large because it is a negative component)
Let H be something. That is, the correction with the edge amount component is for increasing the contrast of the fine line. THRb, THRw
Is set to a value that is easier to extract than fine line pattern matching. However, when the pattern of the pattern matching between the fine line pattern and the fine line pattern 1 is different, the extraction result should be such that the fine line pattern is easier to extract.

In step S41, if they do not match the fine line pattern 1, L (off) is output as output 1 and output 2 of the pattern matching unit B44 (S22), and M
It is determined whether or not FB = 0 (S23), and MFB = 0
If so, the process proceeds to step S36. If MFB is not 0, MFB = MFB-1 is set (S24), and the process proceeds to step S36.

On the other hand, in step S41, the fine line pattern 1
If the state variable MFB is larger than 8, it is determined whether MFB> 8 (S42). If the state variable MFB is larger than 8, L (off) is output as the output 1 of the pattern matching unit B44, and H is output as the output 2. (On) is output (S44),
MFB = MFB-1 is set (S24), and step S24 is executed.
Proceed to 36. If the state variable MFB is not larger than 8, it is determined whether or not MFB = 0 (S43).
If FB = 0, the process proceeds to step S32, where MFB = 0.
Then, L (off) is output to the output 1 and the output 2 of the pattern matching unit 305B (S22), and M is output again.
It is determined whether or not FB = 0 (S23), and MFB = 0
If so, the process proceeds to step S36. If MFB is not 0, MFB = MFB-1 is set (S24), and the process proceeds to step S36.

Further, the fine line pattern matching and the fine line pattern matching 1, the threshold values THRw and THRb may be coded using a magnitude relation. As an example of coding, THRw and THRb of the thin line pattern are THRw and THRb, respectively.
THRw and THRb of the fine line pattern 1 are THRw1 and THRb1 respectively.
Assuming that the magnitude relationship is THRw <THRw1 <THRb1 = THRb, in this case, if coding is not performed, 4 or 3 bits × n lines are obtained. However, if coding is performed as follows, 2 bits × n lines are obtained.

P <THRw → code “0” THRw <P <THRw1 → code “1” THRw1 <P <THRb → code “2” THRb <P → code “3”

Here, P is an edge amount. Since the code is "0" to "3", it can be represented by 2 bits, and when the code is expanded, the reverse process may be performed. In addition, the magnitude relation need not be fixed, but it is of course better to be able to exchange the magnitude relation.

As a result, it is possible to switch between a white background pattern and a halftone dot or color ground pattern using the state variable (MFB), and the state variable can be commonly used.

If the character on the halftone dot can be extracted well (S41 in the flowchart of FIG. 14), the state variable need not be referred to (S43). Is always determined to match). In this manner, characters on a halftone dot and characters on a white background may be separated.

Therefore, in particular, the description of catalog usage
Even if there is a halftone dot in the ruled line as in the specification table, the ruled line can be extracted well. Since the ruled line including the white background is extracted as a character and the character on the halftone dot is determined differently from the character on the white background, the accuracy is further improved compared to the first embodiment. Characters on a white background and characters on a halftone dot can be distinguished to perform different processing.

In FIG. 9, P (i, j + 1),
There are only three propagation directions, P (i + 1, j) and P (i + 1, j-1). In particular, regarding the direction of P (i + 1, J-1), not only -1 but also -2,- It is better to add 3 or the like to eliminate the main scanning directionality of the propagation direction of the state variable.

Further, in a device in which all the image data is stored in the page memory, it goes without saying that the propagation direction of the state variable should be all directions (360 degrees).

FIG. 15 shows detail enhancement processing by unsharp masking. In the figure,
(A) shows a main signal to be processed, (b) shows an unsharp signal, (c) shows an unsharp mask signal, and (d) shows a detail-emphasized signal. Correction is performed based on these edge characteristic examples. Embodiment 1 and Embodiment 2
Then, the Laplacian unit 42 corrects the edge amount by using the unsharp mask signal of FIG.
The correction may be performed using the detail emphasized signal of (d) and other signals.

When the pattern matching section A43 picks up black (outline) on a white background, characters on halftone dots (shaded) are not extracted. The pattern matching unit B44 separately extracts a ruled line on a white background and a ruled line on a halftone dot or a color ground. For example, complicated characters such as "sho" are also extracted by the pattern matching unit B44.

In the first and second embodiments, the upper limit of the state variable is described as 8 (characters on halftone dots) and 16 (characters on a white background). However, the number is not limited to any number.

A method of separating a character on a halftone dot from a character on a white background is shaded in a ruled line to avoid erroneous detection of a character in the ruled line. This is because there is a possibility, and the possibility that a character is nearby decreases as the width of the ruled line increases.

According to the first and second embodiments, small characters, line drawings, characters with a large number of strokes on a white background, and characters on halftone dots can be separately extracted. Since the reflection direction in the sub-scanning direction is one direction, the readout method of the raster scan method is particularly suitable for hardware, and can be easily reflected on image data.

The present invention is an algorithm for detecting the edge of a line drawing. In particular, since a dot peculiar to a printed matter is not detected and removed, a dot such as a generation (copied matter of a copying machine) is included. It is especially effective for manuscripts that do not exist.

Further, the extracted pattern is used to remove small area erroneous judgments by removing isolated points in pixel units by the isolated point removing units A46 and B47. Thereafter, the isolated block removing unit A5 is used.
4, B55, an isolated block is removed in a wide range in a larger block unit (4 × 4), so that erroneous determination can be satisfactorily removed. In this case, the expansion portions A56 and 57
Since the coarse pixel density in the block unit is converted to the original pixel density, there is no coarse pixel in the block unit.

In addition, in the above-mentioned expansion portions A56 and B57,
When simple dilation is performed, only the isolated area becomes large. However, as in the first and second embodiments, when the expansion amount is X, X = M−N, M pixel expansion is performed, and then N pixels are expanded. Since X is smaller than M because pixels are reduced, isolated regions can be connected. Further, since the expansion is performed at a coarse density, in other words, the expansion is performed with the coarse density (block unit), so that the amount of hardware can be reduced.

Further, since the result of the first scan is used in the second and subsequent scans, the line drawing judgment result always matches the color judgment result, so that processing can be performed without variation. The data stored in the page memories A50 and B51 is
Since the data whose processing density has become coarse (roughened) is stored not during the final result but during data processing, the amount of data stored in the memory can be reduced.

A page memory A5 for a line drawing determination result
The determination result at the first scan may be stored in 0 and B51, and may be used after the second scan. In this case, after the second scan, the page memories A50, B51
By reading the characters, the variation in each scan (reading) is only caused by the fluctuation of the edge extraction unit 14C.
The picture determination signal C / P is stabilized.

Instead of storing all the judgment results in the memory, only those having large variations for each scan may be stored in the memory.

In the first and second embodiments described above, the case of four images (full color) has been described. However, the present invention is not limited to this, and the number of images is not limited. For example, it is needless to say that the present invention is also applicable to one image formation (monochrome), for example.

[Operation Sequence of Image Processing Apparatus]
An operation sequence of the image processing apparatus will be described. Such an image processing apparatus includes a memory (also referred to as an image memory) 12.
There are the following modes of operation sequence in combination with The relationship between the operation sequence and the page memory (separation result memory) will be described.

Pressure plate 1 to 1 mode (using memory 12)
Will be described with reference to FIG. In this mode, memory 1
2 is a mode for outputting an image of a size (A4) or less that can be stored in a sheet 2 on paper. In this mode, an image forming operation is performed four times in one scan. First, at the time of scanning of the first color, original image data is taken into an image memory while outputting image data to an image recording unit (also referred to as a printer) 3.
Next, for the second and subsequent colors, image data is output from the image memory to the image recording unit 3.

In this case, in all four image formations, the page memory (separation result memory) is set to the write mode, and the contents of the page memory (separation result memory) are not read. 4
Since both the image forming and the image reading are performed from the image memory, it is not necessary to use the page memory (separation result memory). There is no problem if the page memory (separation result memory) is used for reading after the second scan. The reading start position of the page memory (separation result memory) is initialized for each image formation.

The pressure plate 1 to 1 mode (without using the memory 12) will be described with reference to FIG. This mode is a mode for outputting an image exceeding the size (A4) that can be stored in the memory 12 to a sheet. This is the mode described in the first and second embodiments. In this mode, four images are formed by four scans, and image data is output from an all-color scanner to an image recording unit (also called a printer) 3.

In this case, the page memory (separation result memory) is set to the write mode in the first scan, and the read mode is set after the second scan. As a result, it is less susceptible to variations in the reading of the scanner, and the result of determining characters and photographs is stabilized. The reading start position of the page memory (separation result memory) is initialized for each scan or image formation.

The memory two-chamfer mode (using the memory 12) will be described with reference to FIG. In this mode, the output of an image of a size that can be stored in the memory 12 (A4) is output in two-chamfering. First, 1
At the time of scanning the color, the original image data is taken into each of the RGB image memories while the image data of the scanner is output and imaged to the image recording unit 3. Next, MFGAT for the first color
When A arrives (before the completion of the scanner return), the image data of the first color on the second side is output and imaged from the image memory to the image recording unit 3. Subsequently, the image recording unit 3 outputs and forms the image data of the second color on the first side from the image memory. Then, the image data of the second color of the second surface is output from the image memory to the image recording unit 3. In this mode, four scans
Image formation. The appearance of eight images is due to two chamfers.

In this case, in all four image forming operations, the page memory (separation result memory) is set to the write mode, and the contents of the page memory (separation result memory) are not read.
This is because there is no need to use a page memory (separation result memory) because all four images are read from the image memory. In the case of two-chamfering, a page space is required, and a memory area may be insufficient after the second scan. Therefore, the page memory (separation result memory) is not used.

The memory two-chamfer mode (the memory 12 is not used) will be described with reference to FIG. This mode is two-chamfering when the memory 12 is not used. In the present embodiment, the memory 12 is not necessary because it is present, but a case will be described as having no memory. In this mode, 8 images are scanned and 4 images are formed. Here, two scans are performed per image formation. That is, since there is no image memory, two chamfers are performed on all scanned images.

In this case, at the time of the first scan, the page memory (separation result memory) is set to the write mode, and after the second scan (the remaining seven scans), the read mode is set. As a result, it is less susceptible to variations in the reading of the scanner, and the result of determining characters and photographs is stabilized. The reading start position of the page memory (separation result memory) is initialized every scan.

The ADF 1 to 1 mode (using the memory 12) will be described with reference to FIG. This mode is an application of the two-sided printing. In the normal two-sided printing, the first and second pages are the same original, but this is the case where the first and second pages are different. is there. In the document feeding process in this mode, when the document feeding is performed, the process is terminated, and at this time, the presence / absence of the next document is determined and the process is issued. The scanning process for the image memory in this mode is a two-chamfering process when there is a next original, so a scanning process is issued and the first original image is stored in the RGB image memory.

Next, the copy process will be described. First,
At the scan end of the first color original, the original is replaced.
Next, after the image memory scan process is normally completed, a copy process (printer / scanner synchronous mode) is executed. Subsequently, when the process execution is started, the image data of the first color and the first color is output from the image memory to the image recording unit 3.
Make an image. Then, the image data of the second sheet of the first color is output and imaged from the scanner to the image recording unit 3, and the image data of the second sheet of the second color is output and imaged from the image memory to the image recording unit 3. Next, the image data of the second sheet of the second color is output and imaged from the scanner to the image recording unit 3. Further, the image data of the first sheet of the third color is output and imaged from the image memory to the image recording unit 3, and the image data of the second sheet of the third color is output and imaged from the scanner to the image recording unit 3.

As described above, when the result of the first image formation is stored in the page memory (separation result memory), the page memory (separation result memory) generally secures only the maximum output size. . At this time, the input data amount is the maximum output size + paper interval. Therefore, it is necessary to increase the memory by the width between the sheets. It is possible to prohibit the writing of the page memory (separation result memory) only in the space between sheets, but the control becomes complicated.

Therefore, in order to facilitate control, the page memory (separation result memory) is not used for the path used for image formation from the image memory at the time of image formation. Result memory).
In this case, as the operating condition, the page memory is used only during the second scan and thereafter (when reading the second document). That is, at the time of the first scan, the second scan is started after the start of the first image formation without using the page memory (separation result memory). Then, at the time of the second scan (first scan of the second document), the page memory (separation result memory) is set to the write mode, and data is stored. Subsequently, after the start of the second image, the third scan is started. Further, after the third scan (second scan of the second document), the page memory (separation result memory) is set to the read mode, and the output result of the page memory is used.

This makes it possible to use a page memory (separation result memory) for data from the scanner. Since the control of the page memory only needs to be performed at the time of reading from the scanner, the sequence is fixed and can be easily performed.

In the above-described embodiment, the control may be performed such that the page memory (separation result memory) is used only in the latter half of the image formation in the two-panel trimming.

[0157]

As described above, according to the image processing apparatus of the first aspect, the n-value conversion means for converting analog image data into n-valued (n is a natural number of 2 or more) digital data, A black pixel area determining means for determining an area constituted by a plurality of black pixels around a predetermined black pixel in the digital data as a black pixel area; and a black pixel area determined by the black pixel area determining means. A first state variable setting means for setting a state variable for the corresponding black pixel, a black pixel extracted from the digital data and a corrected black pixel, and a white pixel extracted from the digital data. A thin line candidate extracting means for extracting a pattern as a thin line candidate by using the corrected corrected white pixel, and the thin line candidate is displayed in black using the state variable and the thin line candidate extracted by the thin line candidate extracting means. A judgment / update means for judging whether the character or the picture is a character, and updating the state variable based on the judgment result. Optimum image processing can be performed according to

According to the image processing apparatus of the second aspect, the analog image data is converted into n values (n is a natural number of 2 or more).
N-value conversion means for converting the image data into digital data, and black pixel area determination means for determining an area formed by a plurality of black pixels around a predetermined black pixel in the digital data as a black pixel area A first state variable setting means for setting a state variable for the corresponding black pixel when the black pixel area is determined to be a black pixel area; and a correction for extracting and correcting a black pixel from the digital data A thin line candidate extracting means for extracting a pattern as a thin line candidate using black pixels and a corrected white pixel obtained by correcting a white pixel from the digital data; and a thin line candidate extracted by the state variable and the thin line candidate extracting means. And determining and updating means for determining whether the thin line candidate is a character on a black character, a character on a halftone dot, or a picture, and updating the state variable based on the determination result. Since it was decided to comprise a, black characters on a character on the dot, to distinguish a pattern or the like, it is possible to perform optimal image processing according to each.

According to the image processing apparatus of the third aspect, in the image processing apparatus of the first or second aspect, when a black character on a white background is detected, the second state variable is set. Since the state variable setting means is provided, it is possible to accurately distinguish characters and patterns on a white background even if thin lines and thick lines are mixed.

According to a fourth aspect of the present invention, in the image processing apparatus of the first, second, or third aspect, the thin line candidate extracting means is configured to extract the thin line candidate extraction condition based on the state variable. Is varied to perform pattern extraction, so that the conditions for extracting characters on black and characters on halftone dots can be changed to perform favorable extraction according to the image.

According to a fifth aspect of the present invention, in the image processing apparatus of the first, second, third, or fourth aspect, the image scanning process is performed in a fixed direction, and the reflection direction of the state variable is changed. Is P with respect to the processing pixel P (i, j).
(I, j + 1), P (i + 1, j), P (i + 1, j-
Since the coordinates are set to 1), it is suitable for a raster scan device and can be easily implemented in hardware. That is, it can be easily applied to a copying apparatus or the like.

According to a sixth aspect of the present invention, in the image processing apparatus of the first, second, third, fourth or fifth aspect, the thin line candidate extracting means uses the Laplacian data to generate the black pixel. Since correction of white pixels is performed, fine lines can be satisfactorily extracted.

According to a seventh aspect of the present invention, in the image processing apparatus according to the sixth aspect, the thin line candidate extracting means extracts the Laplacian data as the Laplacian data by scanning speed, magnification, or image rotation. Are made different, so that it is possible to easily cope with scaling, image rotation, and the like.

The image processing apparatus according to the eighth aspect is the image processing apparatus according to any one of the first to seventh aspects, further comprising: determining a character area in the digital data and determining a character area in the digital data. First character area determining means for outputting a rough state determination result, edge extracting means for extracting an edge in the digital data with a pixel density, and a first character area determining means for determining a character area; A second character area determining unit that determines a true character area when the edge is determined to be an edge by the edge extracting means, so that the character area can be accurately detected. Become.

An image processing apparatus according to a ninth aspect of the present invention is the image processing apparatus according to any one of the first to seventh aspects, further comprising determining a character area on a halftone dot in the digital data. A third character area determining means for outputting a determination result in a state rougher than the pixel density; an edge extracting means for extracting an edge in the digital data with the pixel density; And a fourth character area determination means for determining a character area on a true halftone dot when the edge extraction means determines that the character area is an edge. , It is possible to accurately detect a character area on a halftone dot.

[Brief description of the drawings]

FIG. 1 is a schematic block diagram of an image processing apparatus according to a first embodiment.

FIG. 2 is a block diagram of an image processing unit according to the first embodiment.

FIG. 3 is a block diagram of a document recognition unit which is a main part of the present invention.

FIG. 4 is an explanatory diagram showing a symbolized matrix for explaining an operation of detecting a white region or a black region in a Laplacian part of a line drawing recognition unit.

FIG. 5 is a cross-sectional view of a line drawing, and is a conceptual diagram showing a relationship between a white area, a black area, and a threshold.

FIG. 6 is an explanatory diagram showing a relationship between edge amounts (x).

FIG. 7 is an explanatory diagram illustrating an example of determining a thin line based on a difference in state variables in a pattern matching unit of a line drawing recognition unit.

FIG. 8 is an explanatory diagram showing setting of a target pixel in a pattern matching process.

FIG. 9 is an explanatory diagram showing propagation of a state variable MFB in the pattern matching processing of FIG. 12;

FIG. 10 is a flowchart of a pattern matching process.

FIG. 11 is a flowchart of a pattern matching process (pattern matching unit A43).

FIG. 12 is a flowchart of a pattern matching process (pattern matching unit B44) according to the first embodiment.

FIG. 13 is an explanatory diagram illustrating an example of interpolation processing in an expansion unit of a line drawing recognition unit.

FIG. 14 is a flowchart of a pattern matching process (pattern matching unit B44) according to the second embodiment.

FIG. 15 is an explanatory diagram showing detail enhancement processing by unsharp masking.

FIG. 16 is an explanatory diagram illustrating an operation sequence of the image processing apparatus.

FIG. 17 is an explanatory diagram illustrating an operation sequence of the image processing apparatus.

FIG. 18 is an explanatory diagram illustrating an operation sequence of the image processing apparatus.

FIG. 19 is an explanatory diagram illustrating an operation sequence of the image processing apparatus.

FIG. 20 is an explanatory diagram illustrating an operation sequence of the image processing apparatus.

[Explanation of symbols]

 REFERENCE SIGNS LIST 1 document reading unit 2 image processing unit 3 image recording unit 11 RGB γ correction unit 12 memory 13 delay unit 14 document recognition unit 14 a line drawing recognition unit 14 b color determination unit 14 c edge extraction unit 15 RGB filter unit 16 color correction unit 17 UCR unit 18 Doubler 19 CMYBk filter unit 20 CMYBk γ correction unit 21 Gradation processing unit 41 Monochrome conversion unit 42 Laplacian unit 43 Pattern matching unit A 44 Pattern matching unit B 45 Pattern matching unit C 46 Isolated point removing unit A 47 Isolated point removing unit B 48 Pixel density conversion unit A 49 Pixel density conversion unit B 50 Page memory A 51 Page memory B 52 Selector A 53 Selector B 54 Isolated block removal unit A 55 Isolated block removal unit B 56 Expansion unit A 57 Expansion unit B 58 AND operator 59 AND operator 61 White isolated point Removed by unit 62 black isolated point removing section 63 contour extraction unit 64 expands the reduced portion 65 the line delay unit 66 color judging circuit 67 page memory C 68 Selector C

Claims (9)

[Claims] An n-value conversion means for converting analog image data into n-valued (n is a natural number of 2 or more) digital data;
A black pixel area determining unit that determines an area formed by a plurality of black pixels around a predetermined black pixel in the digital data as a black pixel area; If it is determined, first state variable setting means for setting a state variable to the corresponding black pixel, and a black pixel extracted from the digital data and corrected to obtain a corrected black pixel and a white pixel extracted from the digital data. Line candidate extracting means for extracting a pattern as a thin line candidate using the corrected white pixels corrected by the method, and using the state variable and the thin line candidate extracted by the thin line candidate extracting means, the thin line candidate is displayed in black. Judge whether it is a character or a picture, and update the state variable based on the judgment result.
An image processing apparatus comprising: updating means. 2. N-value conversion means for converting analog image data into n-valued (n is a natural number of 2 or more) digital data;
A black pixel area determining unit that determines an area formed by a plurality of black pixels around a predetermined black pixel in the digital data as a black pixel area; If it is determined, first state variable setting means for setting a state variable to the corresponding black pixel, and a black pixel extracted from the digital data and corrected to obtain a corrected black pixel and a white pixel extracted from the digital data. Line candidate extracting means for extracting a pattern as a thin line candidate using the corrected white pixels corrected by the method, and using the state variable and the thin line candidate extracted by the thin line candidate extracting means, the thin line candidate is displayed in black. An image processing apparatus comprising: a determination unit that determines which of a character and a character on a halftone dot is a character or a picture and updates the state variable based on the determination result. 3. The image processing apparatus according to claim 1, further comprising a second state variable setting means for setting said state variable when a black character on a white background is detected. Image processing device. 4. The image processing apparatus according to claim 1, wherein said thin line candidate extracting means performs pattern extraction by changing the thin line candidate extraction conditions based on said state variables. Image processing apparatus. 5. The image processing apparatus according to claim 1, wherein the scanning direction of the image is performed in a fixed direction, and the reflection direction of the state variable is a processing pixel P (i, j).
, P (i, j + 1), P (i + 1, j), P
An image processing apparatus, characterized in that the coordinates are (i + 1, J-1). 6. The image processing apparatus according to claim 1, wherein said thin line candidate extracting means corrects said black pixel and white pixel using Laplacian data. Image processing device. 7. The image processing apparatus according to claim 6, wherein
The image processing apparatus according to claim 1, wherein the thin line candidate extracting means changes weights to be extracted as the Laplacian data depending on a scanning speed, a magnification, or an image rotation. 8. The image processing apparatus according to claim 1, further comprising: determining a character area in the digital data and outputting a determination result of a state rougher than a pixel density. Character area determining means, an edge extracting means for extracting an edge in the digital data with a pixel density, a character area determined by the first character area determining means, and an edge determined by the edge extracting means. And a second character area determining means for determining a true character area when the determination is made. 9. The image processing apparatus according to claim 1, further comprising: determining a character area on a halftone dot in the digital data to determine a determination result of a state rougher than a pixel density. A third character area determining means for outputting, an edge extracting means for extracting an edge in the digital data with a pixel density, and a character area on a halftone dot by the third character area determining means; An image processing apparatus comprising: a fourth character area determination unit that determines a character area on a true halftone dot when an extraction unit determines that the edge is an edge.