JPH07273983A - Image processor - Google Patents

Abstract

PURPOSE:To prevent image quality from deteriorating by smoothly performing black processing for characters, line drawings, etc. CONSTITUTION:On the basis of an RGB signal, a character thickness decision part 114 in a black character decision part 113 decides the thickness of a character/line drawing part in an image. Further, an edge detection part 115 obtains outline information on characters and line drawings and a color saturation decision part 116 obtains color saturation information; when image processing is performed by combining the color saturation information with the outline information, a thickness decision signal is so corrected that the characters and lines continuously vary in thickness.

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 for processing an output image based on the characteristics of the image extracted from the input image.

[0002]

2. Description of the Related Art In recent years, a color printing apparatus that digitally processes color image data and outputs it to a color printer to obtain a color image, or a color original obtained by color-separating and electrically reading a color original Copying a color image with image data printed out on paper, the so-called
There are remarkable developments in color printing systems such as digital color copiers.

Further, with the widespread use of these systems, the demands on the print quality of color images are also increasing, and in particular, the demand for printing black characters and fine black lines in a blacker and sharper manner is increasing. That is, when a black original is color-separated, yellow as a signal for reproducing black,
Magenta, cyan, and black signals are generated, but if you print as they are based on the obtained signals, each color will be reproduced by superimposing four colors, so there will be a slight misalignment between the colors, causing a black fine line to bleed. Occurs, and the original black does not look black, or it looks blurred, which significantly deteriorates the printing quality.

On the other hand, color information such as black and other colors in the image signal and characteristics of spatial frequency such as thin lines and halftone dots are extracted to detect areas such as black characters and color characters. Further, there has been proposed a method of dividing into a halftone image area or a halftone dot image area and performing processing according to each area, and if the area is a black character portion, it is made into a black single color.

Further, by improving the above method and determining the thickness of a character and performing black character processing according to the thickness, it is possible to improve a clear difference in processing at the boundary line of the black character processing. The method is also proposed by the present applicant in Japanese Patent Application No. 5-354528.

[0006]

However, the parameters for the area determination in the above-mentioned conventional example are based on the assumption that the background density of a document, the density of characters, etc. are to some extent.

For this reason, it is difficult to perform the desired black character processing only on the portion of all originals that is suitable for the user's preference.

Further, conventionally, different black character processing could not be performed on a partial area in one screen.

The present invention has been made in view of the above problems, and an object of the present invention is to perform favorable image processing according to the characteristics of the original document and the user's request. More specifically, for example, it is possible to set the area to be subjected to black character processing and the degree of black character processing according to the preference of the user.

It is also possible to change the processing of a line drawing portion of a specific color such as a black character for an arbitrary area in one screen.

[0011]

In order to solve the above-mentioned problems, an image processing apparatus of the present invention is an image processing apparatus for processing image data representing an original image, wherein the characters in the original image are extracted from the image data. · A determination unit for determining a line drawing, a unit for performing a predetermined image processing based on the determination result of the determination unit, and an emphasis on character / line drawing or halftone when extracting characteristics of a document image. It is characterized in that it has a designating means for designating whether or not in multiple stages, a parameter for character / line drawing determination by the determining means based on the designation by the designating means, and a control means for controlling the processing by the processing means.

Further, in an image processing apparatus for processing color image data representing an original image, a detecting means for detecting a line image portion of a specific color from the color image data, a signal generating means for generating predetermined area data, And a control unit that controls a parameter for detection by the detection unit according to the area data.

[0013]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Preferred embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

(First Embodiment) FIG. 1 is a diagram showing a sectional structure of an image processing apparatus according to an embodiment of the present invention. In the figure, reference numeral 201 denotes an image scanner unit, which performs document reading and digital signal processing here. Also,
A printer unit 200 includes an image scanner unit 201.
An image corresponding to the original image read in is printed out on paper in full color.

Next, the document mode in this embodiment will be described.

The operator selects a document mode from the six document modes shown in FIG. 39 according to the document to be copied. The processing for each original mode is as follows.

In the character mode, the recording resolution is 400 dp.
i (dot per inch), the color of the original is discriminated, and black character processing is performed on a portion determined to be black character, and edge emphasis processing is performed on a region other than black. By this process, it is possible to reproduce fine characters in detail by increasing the recording resolution, and it is possible to reproduce black characters with clear color blur by recording only black toner for black characters.

In the map mode, edge enhancement processing is performed,
Recording is performed at a recording resolution of 400 dpi. Also, a masking UCR coefficient with a stronger UCR is used. By this processing, it is possible to reproduce an image with high resolution for a map document having many fine characters and lines. Further, since the area separation is not performed, an output image can be obtained in which the quality is not deteriorated due to the erroneous determination that occurs during the area separation. Further, by increasing the UCR, it becomes possible to record black characters existing in the original document at a ratio in which the black toner is increased while suppressing the color toner as much as possible.

In the photographic paper photograph mode, edge enhancement for photographic paper photographs is performed and recording is performed at a recording resolution of 200 dpi.
By this processing, it is possible to output an image having high gradation and sharpness of the image being emphasized and having a heap.

In the print photograph mode, smoothing processing is performed to suppress the occurrence of moire, edge enhancement is performed, and recording is performed at a recording resolution of 200 dpi. By this processing, it is possible to output an image with high gradation and enhanced image sharpness without generating moire.

In the character print photograph mode, the character area and the print photograph area are automatically identified, and the area determined to be the character area is processed for characters and the area determined to be the print picture area is processed for the print photograph. Is given.

In the character photographic paper photograph mode, the character area automatically identifies the photographic paper photographic area, the area determined to be the character area is processed for characters, and the area determined to be the photographic paper photograph area is processed to the photographic paper photograph. Processing is performed.

The user selects the entire manuscript as shown in FIGS.
Not only can the original mode be selected from the operation unit 101 shown in FIG.
By setting a plurality of original mode areas on the original using, it is possible to set different original modes for each area. In the above mode setting, the CPU 102 sets the LU
This can be realized by controlling the output of T117.

Then, the user presses the copy start key to start the copy operation.

In the image scanner section 201, the original 204 placed on the original platen glass (platen) 203 by the original pressing plate 202 is irradiated with the light of the halogen lamp 205. The reflected light from the original 204 is reflected by the mirror 206,
It is guided to 207 and forms an image on a three-line sensor (hereinafter referred to as CCD) 210 by a lens 208. An infrared cut filter 231 is provided on the lens 208.

The CCD 210 color-separates the light information from the original 204, reads the red (R), green (G), and blue (B) components of the full-color information from the color information, and then sends it to the signal processing unit 209. Each color component reading sensor array of the CCD 210 is composed of 5000 pixels. As a result, 297 mm in the lateral direction of the A3 size document, which is the largest size of the documents placed on the platen glass 203, is read at a resolution of 400 dpi.

The halogen lamp 205 and the mirror 20
6 is the velocity v, and the mirror 207 is (1/2) v,
The entire surface of the original 204 is scanned by mechanically moving in a direction (hereinafter, referred to as a sub-scanning direction) perpendicular to an electric scanning direction (hereinafter, referred to as a main scanning direction) of the line sensor 210.

The standard white plate 211 is the R, G, B sensor 2
The correction data of the read data in 10-1 to 210-3 is generated. The standard white plate 211 exhibits a substantially uniform reflection characteristic with visible light, and has a white color in the visible. Here, using this standard white plate 211, R,
The output data from the G and B sensors 210-1 to 210-3 are corrected.

In the image signal processing unit 209, the read signal is electrically processed and decomposed into magenta (M), cyan (C), yellow (Y) and black (Bk) components, It is sent to the printer unit 200. Further, one component of M, C, Y, and Bk is sent to the printer unit 200 (field-sequential image formation) per one document scanning (scan) in the image scanner unit 201, and a total of four document scanning is performed. This completes the printout for one sheet.

In the printer section 200, the M, C, Y and Bk image signals from the image scanner section 201 are sent to the laser driver 212. The laser driver 212 is
The semiconductor laser 213 is modulated and driven according to the image signal. Then, the laser light scans the photosensitive drum 217 via the polygon mirror 214, the f-θ lens 215, and the mirror 216.

The developing device is composed of a magenta developing device 219, a cyan developing device 220, a yellow developing device 221, and a black developing device 222, and these four developing devices are alternately in contact with the photosensitive drum 217, and are placed on the photosensitive drum 217. The formed M, C, Y, and Bk electrostatic latent images are developed with corresponding toners. In addition, the transfer drum 223 is the paper cassette 2
24, or the paper fed from the paper cassette 225 is wound around the transfer drum 223, and the toner image developed on the photosensitive drum 217 is transferred onto the paper.

In this way, the toner images for the four colors M, C, Y, and Bk are sequentially transferred, and then the sheet passes through the fixing unit 226 and is ejected.

Next, the image scanner unit 201 according to this embodiment will be described in detail.

FIG. 2 is a diagram showing the external structure of the CCD 210. In the figure, 210-1 is an example of a light receiving element (photo sensor) for reading red light (R),
Reference numerals 0-2 and 210-3 denote light receiving element arrays for reading the wavelength components of the green light (G) and the blue light (B), respectively.
These R, G, B sensors 210-1 to 210-3
Has an opening of 10 μm in the main scanning direction and the sub scanning direction.

The above three light receiving element arrays having different optical characteristics are arranged on the same silicon chip so that the R, G and B sensors are arranged in parallel with each other so as to read the same line of the original. Takes a monolithic structure. Further, by using the CCD having such a configuration, the optical system such as a lens is shared in each color separation reading, and thereby it is possible to simplify the optical adjustment for each color of R, G, and B. .

FIG. 3 is a sectional view of the image scanner unit 201 taken along the dotted line aa 'shown in FIG.
As shown in the figure, a photosensor 210-1 for reading R color and photosensors 210-2, 210-3 for reading visible information of G and B are provided on a silicon substrate 210-5.
Are arranged.

On the R color photosensor 210-1, an R filter 210 that transmits the R color wavelength component of visible light is transmitted.
-7 is placed. Similarly, the G color photo sensor 210
-2, the G filter 210-8 is arranged, and the B color photosensor 210-3 is arranged on the B filter 210-9. Incidentally, 210-6 is a flattening layer made of a transparent organic film.

FIG. 4 is an enlarged view of the light-receiving element indicated by reference numeral B in FIG. As shown in FIG. 4, each sensor has a length of 10 μm per pixel in the main scanning direction. Each sensor has 5000 pixels in the main scanning direction so that the A3 size document can be read in the lateral direction (length 297 mm) at a resolution of 400 dpi as described above. The distance between the lines of the R, G, and B sensors is 80 μm, and each line is separated by 8 lines for a resolution of 400 dpi in the sub-scanning direction.

Next, the density reproduction method in the printer section of the image processing apparatus according to this embodiment will be described.

In this embodiment, in order to reproduce the density of the printer, PWM (pulse width modulation), which is well known in the past, is used.
The method controls the lighting time of the semiconductor laser 213 according to the image density signal. As a result, an electrostatic latent image having a potential according to the laser lighting time is formed on the photosensitive drum 217. Then, the developing units 219 to 222 develop the latent image with toner in an amount corresponding to the potential of the electrostatic latent image, thereby reproducing the density.

FIG. 5 is a timing chart showing the control operation of density reproduction in the printer section according to this embodiment. Reference numeral 10201 is a printer pixel clock, which is 4
This corresponds to a resolution of 00 dpi (dot per inch). The clock is generated by the laser driver 212. Also, a 400-line (line per inch) triangular wave 10202 is generated in synchronization with the printer pixel clock 10201. The cycle of the triangular wave 10202 of 400 lines is the same as the cycle of the pixel clock 10201.

40 sent from the image signal processing unit 209
256 gradations (8 bits) of M, C, with 0 dpi resolution
The Y and Bk image data and the 200-line / 400-line switching signal are transmitted in synchronization with the CLOCK signal, but the laser driver 212 synchronizes with the printer pixel clock 10201 by a FIFO memory (not shown). Be seen. This 8-bit digital image data is D /
An analog image signal 10203 by an A converter (not shown)
Is converted to. Then, the above 400-line triangular wave 1020
Analogly compared with 2, and as a result, 400-wire PW
M output 10204 is generated.

Digital pixel data is 00H (H is 16
From 400) to FFH, and the 400-line PWM output 10204 has a pulse width corresponding to these values. Further, one cycle of 400-line PWM output is 63.5 μm on the photosensitive drum.

In addition to the 400-line triangular wave, the laser driver 212 also produces a 200-line triangular wave 10205 having a period twice that of the 400-line triangular wave in synchronization with the printer pixel clock 10201. Then, the 200-line triangular waves 10205 and 400
A 200-line PWM output signal 10206 is generated by comparing with the analog image signal 10203 of dpi. The 200-line PWM output signal 10206 forms a latent image on the photosensitive drum at a period of 127 μm, as shown in FIG.

In the density reproduction with 200 lines and the density reproduction with 400 lines, since the minimum unit for density reproduction of 200 lines is 127 μm, which is twice that of 400 lines, gradation reproducibility is good. However, in terms of resolution, 400 lines that reproduce the density in units of 63.5 μm are more suitable for high-resolution image recording. As described above, the 200-line PWM recording is suitable for gradation reproduction, and the 400-line PWM recording is superior in terms of resolution.
And 400 lines of PWM are switched.

The signal for performing the above switching is the 200-line / 400-line switching signal 10207 shown in FIG. 5, which is sent from the image signal processing unit 209 to the laser driver 212 in pixel units in synchronization with the 400 dpi image signal. Is entered. This 200-line / 400-line switching signal is logical Lo
In the case of w (hereinafter referred to as L level), P of 400 lines
When the WM output is selected and it is a logic high (hereinafter, referred to as H level), the PWM output of 200 lines is selected.

Next, the image signal processing unit 209 will be described.

FIG. 6 is a block diagram showing the flow of image signals in the image signal processing section 209 of the image scanner section 201 according to this embodiment. As shown in the figure, CCD
The image signal output from 210 is input to the analog signal processing unit 101, where gain adjustment and offset adjustment are performed, and then the A / D converter 102 performs 8-bit digital image signals R1, G1, and B1 for each color signal. Is converted to. Then, the shading correction unit 103 inputs the known shading correction using the read signal of the standard color plate 211 for each color.

The clock generator 121 generates a clock for each pixel. In addition, the main scanning address counter 12
In 2, the clock from the clock generator 121 is counted to generate a pixel address output for one line. And
The decoder 123 decodes the main scanning address from the main scanning address counter 122 to obtain a CCD driving signal for each line such as a shift pulse and a reset pulse, and a CCD.
The VE signal and the line synchronization signal HSYNC which represent the effective area in the 1-line read signal from are generated. The main scanning address counter 122 is cleared by the HSYNC signal and starts counting the main scanning address of the next line.

As shown in FIG. 2, since the light receiving portions 210-1, 210-2, 210-3 of the CCD 210 are arranged at a predetermined distance from each other, the line delay circuits 104, 105 shown in FIG. At, the spatial deviation in the sub-scanning direction is corrected. Specifically, the R and G signals are line-delayed in the sub-scanning direction with respect to the B signal in the sub-scanning direction.
Match the signal.

The input masking section 106 is a CCD 210.
R, G, B filters 210-7, 210-8, 21
The read color space determined by the spectral characteristics of 0-9
Is a part that is converted into the standard color space of, and the matrix operation as in the following equation is performed.

[0052]

[Outer 1]

Light / concentration converter (LOG converter) 107
Is a lookup table ROM, R4,
The luminance signals of G4 and B4 are converted into density signals of C0, M0 and Y0. The line delay memory 108 delays the image signals C0, M0, and Y0 by a line delay from the R4, G4, and B4 signals to the generated determination signals such as UCR, FILTER, and SEN in the black character determination unit 113 described later. Let As a result, the C1, M1, and Y1 image signals for the same pixel and the UCR of the black character determination signal are simultaneously input to the masking UCR circuit 109.

The masking and UCR circuit 109 receives the black signal (B) by the input three primary color signals of Y1, M1, and C1.
k) is extracted, and the calculation for correcting the color turbidity of the recording color material in the printer 212 is performed to obtain Y2, M2, C2, and B.
The k2 signal is output in a frame-sequential manner with a predetermined number of bits (8 bits) each time the image scanner unit 201 performs each reading operation.

The main-scanning scaling circuit 110 performs enlargement / reduction processing of the image signal and the black character determination signal in the main-scanning direction by a known interpolation calculation. Further, the spatial filter processing unit (output filter) 111, as described later, based on the 2-bit FILTER signal from the LUT 117, edge enhancement,
Switches smoothing processing.

Thus processed M4, C4, Y4,
The Bk4 frame-sequential image signal and the SEN signal, which is a switching signal of 200 lines / 400 lines, are supplied to the laser driver 2 described above.
Then, the density is recorded by PWM in the printer unit 200.

FIG. 7 shows the image signal processing unit 209 shown in FIG.
3 is a diagram showing the timing of each control signal in FIG. In the figure, the VSYNC signal is an image effective section signal in the sub-scanning direction, and in the section of logic "1", image reading (scanning) is performed, and (M), (C),
The output signals of (Y) and (Bk) are formed. Further, the VE signal is an image effective section signal in the main scanning direction, and is used mainly for the line counting control of the line delay by timing the main scanning start position in the section of logic "1". Then, the CLOCK signal is a pixel synchronization signal, and "0" →
Image data is transferred at the rising timing of "1",
The laser driver 2 supplies the signal to the signal processing units such as the A / D converter 102 and the black character determination unit 113.
It is used to transmit an image signal and a switching signal of 200 lines / 400 lines to 12.

Next, the black character processing executed when the original mode is the character mode, the character print photo mode, or the character photographic paper photo mode will be described.

(Description of Edge Detection Unit) As described above, the signal R masked by the input masking unit 106 is converted.
4, G4, and B4 are input to the edge detection circuit 115 of the black character determination unit 113, and the luminance signal Y is calculated according to the following formula. 8 is a block diagram showing the internal configuration of the edge detection circuit 115, and FIG. 9 is a luminance calculation circuit 25.
It is a figure which shows the detailed structure of 0.

Y = 0.25R + 0.5G + 0.25B (2)

In FIG. 9, the input color signals R, G,
B is multiplied by coefficients 0.25, 0.5, and 0.25 in multipliers 301, 302, and 303, and then added in adders 304 and 305, respectively, according to the above equation (2). The luminance signal Y is calculated.

The luminance signal Y is the FIFO 40 shown in FIG.
1 to 402, the line is extended to three lines delayed by one line, and known Laplacian filters 403 to 40
Can be multiplied by 6. Then, of the four directions shown in the figure, the direction in which the absolute value a of the edge amount, which is the output of the filter, takes the minimum value is obtained, and this direction is defined as the edge min direction. This is performed by the edge min direction detection unit 251 shown in FIG.

Next, the edge min direction smoothing unit 2
At 52, smoothing processing is applied to the edge min direction obtained by the edge min direction detection unit 251. By this processing, only the direction with the largest edge component can be saved and the other directions can be smoothed.

That is, a halftone dot component having a large edge component in a plurality of directions is smoothed by the above processing and its characteristics are reduced, while a character having an edge component in only one direction is present. / For thin lines, the effect that the characteristics are preserved is enhanced. By repeating this process, if necessary, the line and halftone dot components can be more effectively separated, and characters existing in halftone dots that could not be detected by the conventional edge detection method. It is also possible to detect the component.

After that, in the edge detecting section 253 shown in FIG. 8, an appropriate threshold value th_edg is applied to the signal directly input from the luminance calculating circuit 250 to the edge detecting section 253.
Those below e are removed, and those above th_edge are output as logic "1". Furthermore, the edge mi
The signal that has passed through the n-direction smoothing unit 252 is subjected to the Laplacian filter described above, and compared with the threshold value th_edge2 having different absolute values of the edge amounts,
The output value is coded according to the rule described later. In this way, by using the edges having two kinds of characteristics properly and using the edges that do not pass through the smoothing portion in the edge min direction for the characters in the white background, the edges can be detected even in the finer details of the characters. Is possible,
Conversely, for characters in halftone dots, it is possible to detect only characters and lines without detecting halftone dot components by performing edge detection using edges that have passed through the edge min direction smoothing section. become.

FIG. 11 is a diagram showing an example of edge detection. From the image data (a) relating to the luminance data Y,
The edge detection signal (b) is generated.

The edge detection unit 115 further outputs the result of the determination made by th_edge among the above determination signals by 7 ×.
An output signal “edge” from the edge detection unit 115 is represented by seven codes including a signal expanded in a block size of 7, 5 × 5, 3 × 3, no expansion, and a signal determined by th_edge2. (3 bits). Here, the signal expansion means ORing the signals of all the pixels in the block.

(Explanation of Saturation Determining Section) FIG. 12 is a block diagram showing a detailed configuration of the saturation determining circuit 116 constituting the black character determining section 113. Here, with respect to the input color signals R4, G4, B4, the maximum value max (R, G,
B) and the minimum value min (R, G, B) are extracted, respectively. Then, in the LUT (look-up table) 603 in the next stage, threshold values Cr_BK, Cr_COL, Cr_W for dividing the data into areas as shown in FIG.
Thus, the saturation signal Cr is generated.

The output signal "col" from the saturation determining section 116 shown in FIG. 6 has the data Bk shown in FIG.
2-bit code is black when entering, intermediate (color between black) when entering GRY, color when entering COL, and white when entering W. Expressed.

(Explanation of Character Thickness Determining Section) FIG. 14 is a block diagram showing the configuration of the character thickness determining circuit 114 constituting the black character determining section 113.

In FIG. 14, the input masking circuit 10
The red signal data R4, the green signal G4, and the blue signal B4 which are outputs from 6 are input to the minimum detection unit 2011. In this minimum value detection unit 2011, the input RG
The minimum value MIN _RGB of the B signal is obtained. Next, MIN _RGB is input to the average value detection unit 2012, and the average value AVE5 of MIN _RGB of 5 pixels × 5 pixels in the vicinity of the target pixel,
Similarly, the average value AVE of MIN _RGB of neighboring 3 pixels x 3 pixels
Ask for 3.

AVE5 is sent to the character / halftone detection unit 2013.
And AVE3 are input. Here, by detecting the density of the target pixel for each pixel and the amount of change between the target pixel and the average density in the vicinity thereof, the target pixel is a part of a character or halftone region. Determine if there is.

FIG. 15 shows a character / halftone detection circuit 2013.
3 is a block diagram showing the internal configuration of FIG. In the character / halftone detection circuit, as shown in FIG.
At 0, add an appropriate offset value OFST1 to AVE3,
The comparator 2031 compares the value with AVE5. Further, in the comparator 2032, the adder 2030
From the output from the appropriate limit value LIM1. And the output value from each comparator is
It is input to the OR circuit 2033.

In the OR circuit 2033, when AVE3 + OFST <AVE5 ... (3) or AVE3 + OFST1 <LIM1 ... (4), the output signal BINGRA becomes logic "H".
In other words, if there is a density change near the pixel of interest (character edge portion) or if the vicinity of the pixel of interest has a density equal to or higher than a certain value by this character / halftone detection circuit (inside and inside the character). The character / halftone area signal BINGRA becomes a logic "H" in the key section.

On the other hand, in the halftone dot area detection unit 2014 whose detailed configuration is shown in FIG. 16, the MIN _RGB detected by the minimum value detection circuit 2011 is first detected in order to detect the halftone dot area.
And an appropriate offset value OFST2 in the adder 2040
And the result is added to the AVE5 by the comparator 2041.
Compare with. Further, the comparator 2042 compares the output from the adder 2040 with an appropriate limit value LIM2. Then, each output value is the OR circuit 204.
3 and the output signal BINAM from the OR circuit 2043 when MIN _RGB + OFST2 <AVE5 (5) or MIN _RGB + OFST2 <LIM2 (6)
I becomes logic "H". Then, using this BINAMI signal, the edge direction detection circuit 2044 obtains the edge direction of each pixel.

FIG. 17 is a diagram showing the rules for edge direction detection in the edge direction detection circuit 2044. That is, when the eight pixels near the target pixel satisfy any of the conditions (0) to (3) shown in FIG. 17, the edge direction signal DIR
Any of 0 to 3 bits of AMI becomes logic "H".

Further, the counter edge detection circuit 204 of the next stage
In 5, the edges that face each other are detected in the area of 5 pixels × 5 pixels surrounding the pixel of interest. Therefore, FIG.
As shown in, the rule of the opposite edge detection in the coordinate system in which the DIRAMI signal of the target pixel is A33 is shown below. That is, (1) A11, A21, A31, A4
Bit 0 of any one of A1, A51, A22, A32, A42, and A33 is "H", and A33, A24, and A3
4, A44, A15, A25, A35, A45, A55
Bit 1 of any of the two is "H", (2) A11, A2
1, A31, A41, A51, A22, A32, A4
2, bit 1 of either A33 is "H", and A3
3, A24, A34, A44, A15, A25, A3
5, bit 0 of A45 or A55 is "H",
(3) A11, A12, A13, A14, A15, A2
2, A23, A24, A33, any bit 2 is "H", and A33, A42, A43, A44, A5
Bit 3 of any one of A1, A52, A53, A54, A55 is "H", (4) A11, A12, A13, A1
4, A15, A22, A23, A24, A33, any bit 3 is "H", and A33, A42, A4
3, A44, A51, A52, A53, A54, A55
Any one of bit 2 is “H”, and the above (1) to
When any of the conditions in (4) are met, EAAM
I is set to "H" (when the opposite edge is detected by the opposite edge detection circuit 2045, the opposite edge signal EAAM
I becomes "H").

The expansion circuit 2046 expands the EAAMI signal by 3 pixels × 4 pixels to generate 3 pixels in the vicinity of the pixel of interest.
If there is a pixel of which the EAAMI is “H” in the pixel × 4 pixels, the EAAMI signal of the target pixel is replaced with “H”. Furthermore, the contraction circuit 2047 and the expansion circuit 2048 are used to remove the detection result isolated in the area of 5 pixels × 5 pixels, and the output signal EBAMI is obtained. Here, the contraction circuit 2047 is a circuit that outputs "H" only when all the input signals are "H".

Next, the counting unit 2049 counts the number of pixels whose output signal EBAMI of the expansion circuit 2048 is "H" within a window having an appropriate size. In this embodiment, an area of 5 pixels × 64 pixels including a target pixel is referred to. The window shape is shown in FIG.

In FIG. 19, the sampling points in the window are 9 points at intervals of 4 pixels in the main scanning direction and 5 in the sub scanning direction.
There are 45 points in total for the line. For one pixel of interest,
By moving the windows in the main scanning direction, nine windows (1) to (9) in FIG. 19 are prepared. That is, 5 pixels x 64 with the pixel of interest as the center
This refers to the pixel area. Then, EBAMI is counted in each window, and EBAM
An appropriate threshold value th_count is the number at which I becomes “H”.
If t is exceeded, the halftone dot area detection unit 2014 of FIG.
Outputs the halftone dot area signal AMI as logic "H".

By the processing in the halftone dot area detection circuit 2014, the halftone dot image detected as a set of isolated points in the BINGRA signal can be detected as the area signal. The detected character / halftone area signal BINGRA and halftone dot area signal AMI are shown in FIG.
Is ORed in the OR circuit 2015, and as a result, the binarized signal PICT of the input image is generated. This PI
The CT signal is input to the area size determination circuit 2016, where the area size of the binarized signal is determined.

Here, a set of isolated points will be briefly described.

The above-described image area determination is performed on a binary image by binarizing the image with a certain density. At this time, the dots and lines are
Areas with letters and areas are judged to be halftone. But,
If the halftone dot image is simply binarized, a dot aggregate that is a constituent element of the halftone dot is generated.

Therefore, by determining whether or not a set of isolated points exists in a region having a certain area,
It is determined whether or not the dot is a halftone dot image. That is, if there are a considerable number of dots in a certain area, that area is a halftone dot image, and if there is no dot around the pixel even if the pixel of interest is part of the dot, that pixel of interest is Is determined to be a part of.

FIG. 20 shows the area size determination circuit 2016.
3 is a block diagram showing the internal configuration of FIG. In the circuit shown in the figure, there are a plurality of pairs of contraction circuits 2081 and expansion circuits 2082, and the sizes of the regions to be referred to are different from each other. The input PICT signal is line-delayed according to the size of the contracting circuit and then input to the contracting circuit 2081. In this embodiment, from 23 pixels × 23 pixels, 35
Seven types of contraction circuits with a size of up to 35 pixels are prepared.

The signal output from the contraction circuit 2081 is line-delayed and then input to the expansion circuit 2082. In this embodiment, 7 of a size from 27 pixels × 27 pixels to 39 pixels × 39 pixels is corresponded to the output of the contraction circuit.
Different kinds of expansion circuits are prepared, and the output signal PICT_FH from each expansion circuit is obtained.

Regarding the output signal PICT_FH, when the pixel of interest is a part of a character, the output of PICT_FH is determined by the thickness of the character. This state is shown in FIG. For example, when the PICT signal exists in a band having a width of 26 pixels, if the contraction of a size larger than 27 × 27 is performed, all the outputs become 0, and 25 × 2.
When contraction of a size smaller than 5 is performed and then expansion according to each size is performed, a band-shaped output signal PICT_FH having a width of 30 pixels is obtained.

Then, by inputting these outputs PICT_FH to the encoder 2083, the image area signal ZONE_P to which the pixel of interest belongs is obtained. Note that FIG.
FIG. 9 is a diagram showing an encoding rule of the encoder 2083.

By such processing, a photographic image and a halftone dot image in which the PICH signal is "H" in a wide area are
A character or line image defined as a region 7 (maximum value) and having an area size smaller than the maximum value (thin) is defined as a multi-valued image region corresponding to its size (thickness). In this embodiment, the ZONE signal has 3 bits, and the thickness of the character is expressed in 8 steps. The thinnest character is 0, and the thickest character (including the area other than the character) is 7.

The ZONE correction unit 2084 shown in FIG.
As shown in FIG. 21, an average value calculation unit 21 to which the ZONE_P signal line-delayed by a plurality of FIFOs is input
10 where the average value of 10 pixels × 10 pixels is calculated. The ZONE_P signal has a larger value as the character is thicker, and has a smaller signal value as the character is thinner, so that the output of the average value calculation unit is the same as the corrected Z value.
It becomes an ONE signal.

Here, it is desirable that the block size used for the correction is determined according to the size of the block size for determining the thickness of the character. And this correction Z
By performing the subsequent processing using the ONE signal, the judgment of the thickness changes smoothly even in the portion where the thickness of the character / line changes abruptly, and the deterioration of the image quality due to the change in the black character processing occurs. , More improved.

Here, as described above, the area where the ZONE signal is in stage 7 can be regarded as the halftone area.
Therefore, by utilizing this, it is possible to distinguish the character / line existing in the halftone dot or halftone area from other area characters / lines based on the ZONE signal and the edge signal. The method will be described below.

FIG. 24 is a diagram showing an algorithm for detecting a character in a halftone dot / halftone. Here, first, with respect to the above-mentioned PICT signal, the expansion process is performed in the portion indicated by reference numeral 2111 in 5 × 5 blocks. By this processing, for the halftone dot area that is likely to be incompletely detected,
The detection area is corrected.

Next, with respect to this output signal, reference numeral 211
In the part indicated by 2, contraction processing of an 11 × 11 block is performed. Signal FCH obtained by these processes
Is a signal contracted by 3 pixels from the PICT signal.

FIG. 25 is a diagram concretely showing a state of processing by the above algorithm. In the figure, by combining the FCH signal, the ZONE signal, and the edge signal, the edge in the white background and the edge in the halftone / halftone can be distinguished, and the halftone component is emphasized even in the halftone image. It is shown that the black character processing can be performed without causing the black character processing and without processing the portion such as the edge of the photograph where the black character processing is unnecessary.

(Explanation of LUT) Next, the LUT 117 which constitutes the black character determining portion 113 shown in FIG. 6 will be explained.

The LUT 117 inputs the signals judged by the character thickness judging section 114, the edge detecting section 115, and the saturation judging section 116 of FIG. 6, and inputs “ucr” according to the table as shown in FIG. , "Filter", "sen" output signals for each processing. These are signals for controlling the masking UCR coefficient, the spatial filter coefficient, and the printer resolution, respectively.

In the table shown in FIG. 26, the meaning of each signal and its value is as follows: edge-0: threshold value th_edg.
e is not judged as an edge 1: No expansion is judged by the threshold value th_edge 2: 3 × 3 expansion is judged by the threshold value th_edge 3: 5 × 5 expansion is judged by the threshold value th_edge 4: Threshold 7 × 7 expansion determined by the value th_edge 5: not determined as an edge by the threshold value th_edge2 6: no expansion determined by the threshold value th_edge2 Sen-0: 200 line, 1: 400 line filter-0: smoothing, 1: Strong edge enhancement,
2: Medium edge enhancement 3: Weak edge enhancement ucr-0 to 7: black to black little FCH: 0: image edge, 1: not image edge Further, the table shown in FIG. 1) Multi-valued black character processing is possible according to the thickness of the character. (2) Since a plurality of edge area ranges are prepared, the black character processing area can be selected according to the thickness of the character. In this embodiment, the widest area is processed even for the thinnest character. (3) The black character processing is performed by making a difference in the degree of processing between the edge of the character and the inside of the character to obtain a smoother black amount. (4) Characters in halftone / halftone are processed separately from those in white background. (5) Character edge, inside of character, halftone / halftone image , The coefficient of the spatial filter is changed. Also,
The coefficient is changed according to the thickness even for the character edge. (6) The resolution of the printer is changed according to the thickness of the character. (7) For the color character, all are the same as the black character except the masking UCR coefficient. It is to do processing.

Needless to say, not only the processing in this embodiment, but various processing methods by various combinations of input signals can be considered.

On the other hand, in the masking UCR processing circuit 109, the UCR control signal ucr output from the LUT 117 is output.
Thus, the black signal Bk is generated and the output masking is performed.

FIG. 28 shows the masking UCR arithmetic expression.

First, the minimum value MIN of C1, M1 and Y1.
_CMY is obtained, and Bk1 is obtained from the equation (2101). Next, 4 × 8 masking is performed according to equation (2102), and C2, M2, Y2, and Bk2 are output. This formula (2
102), the coefficients m11 to m84 are masking coefficients determined by the printer used, and the coefficients k11 to k8.
4 is a UCR coefficient determined by the UCR signal.

For the halftone dot / halftone image (ZONE signal is 7), the UCR coefficients are all 1.0, but for the thinnest character (ZONE signal is 0), Bk single color is output. To set the UCR coefficient. Further, for the intermediate thickness, the UCR coefficient is determined so that the change in the tint corresponding to the thickness is smoothly connected, and the amount of Bk is controlled.

In the spatial filter processing section 111, 5
Two filters of pixels × 5 pixels are prepared, and the output signal of the first filter is connected to the input of the second filter. There are four filter coefficients, smoothing 1, smoothing 2, edge enhancement 1 and edge enhancement 2, and L
The coefficient is switched for each pixel by a filter signal from the UT 117. Further, by using two filters, edge enhancement after smoothing is performed to realize edge enhancement with reduced moire, and by combining two types of edge enhancement coefficients, a higher quality image is output. It is possible.

Here, when it is desired to preferentially process the characters in the original, the adjustment of the character / photo separation level shown in FIG. 29 is given priority to the character, and when it is desired to preferentially process the photo in the original, the character / photo separation level is adjusted. By adjusting the photo priority destination, it is possible to provide an output according to the user's preference. In this embodiment, it is possible to adjust in four steps in both directions.

Next, the content of each processing parameter that changes depending on the adjustment of the character / photo separation level will be described.

By adjusting the character / photo separation level, each judgment parameter of the character / photo area and each parameter for adjusting the degree of character processing are adjusted at the same time.

FIG. 30 shows the judgment parameters for the adjusted character / photo area. The parameters to be adjusted are the edge detection threshold th_edge in the white background, the edge detection threshold th_edge2 in the halftone dot, and the saturation determination threshold Cr.
_Bk, Cr_col, Cr_w, threshold values LIM1 and OFST for generating character / halftone area signal BINGRA
1, threshold value LIM2 for determining a halftone dot area, th_c
It is the window size of the contraction circuit for determining the out and thickness.

By changing the threshold of edge detection, it is possible to detect even thin characters and fine characters. With character priority, the edge threshold is lowered to detect even thin and fine characters. If the priority is set on the photo, the threshold of the edge is increased to make it difficult to detect the character. Further, in this embodiment, since the threshold value for detecting the edge in the white background and the threshold value for detecting the edge in the halftone dot are independent, the detection accuracy of the edge in the white background remains the same. It is possible to change the inner edge detection accuracy, or change the edge detection accuracy in the white background while keeping the edge detection accuracy in the halftone dots. In this embodiment, when the step of giving priority to the photograph is enlarged, the character processing in the halftone dot is not performed.

LIM1, OFST1, LIM2, th_
By changing the count, it is possible to detect characters according to the density of the background of the original. With character priority, even if the original has a high background density, the original is regarded as a white background and the characters are detected, and the photo priority is applied. Then, for a document having a background density, character processing is performed by regarding the document as a halftone image.

By changing the saturation judgment threshold value,
It is possible to change the condition for performing the processing to output with black monochrome toner, and it becomes easy to process even pencil writing and faint characters into black monochromatic color with character priority, and black letters and lines with low saturation are colored with photo priority. It becomes difficult for monochromatic processing.

By changing the window size of the contraction circuit for determining the thickness, the width of the line on which the black character processing is performed can be changed. In the case of character priority, the window size of the contraction circuit for determining the thickness is increased to facilitate black character processing even for wide lines and thick characters. By making it small, it is difficult to apply black character processing to wide lines and thick characters.

31 to 38 show the contents of the LUT corresponding to the degree of character processing adjusted by adjusting the character / photo separation level using the operation unit 101 shown in FIG.

When the character priority is set, the processing area at the edge portion is further expanded, and the UCR control is also performed on the inside of the character, whereby the reproduction of the black character becomes smoother. In addition, the control to output the characters in the halftone dots more clearly is strengthened.

When the photograph priority is set, the area processed for the edge is narrowed and the processed line width is narrowed. Further, when the degree of priority increases, the character processing in the halftone dot is not performed, so that the quality of the photograph is prioritized.

As described above, according to this embodiment,
Judgment parameter and processing for black character processing by adjusting the character / photo separation level when determining the thickness of the character / line drawing part in the image and performing image processing by combining outline information and saturation information of the character / line drawing. By simultaneously changing the degree of, the thickness of the character to be processed and the degree of processing change, and it becomes possible to reproduce the image according to the user's preference.

(Second Embodiment) In addition to the configuration of the first embodiment, the present embodiment is such that the determination result parameter by the LUT 117 can be changed for each area.

As shown in FIG. 40, in the present embodiment, the digitizer 100 is provided so that the mode setting of FIG. 39 in the first embodiment can be performed for each area designated by the digitizer 100.

For example, the black character processing shown in the above-described first embodiment is performed on the pixel of which the area data from the digitizer 100 is "0", and the pixel of "1" is uc.
By performing the time-division processing in which the fixed values of r7, filter0, and sen0 are performed, it is possible to prevent the normal black character processing from being performed on a partial area.

Further, the ZONE signals "0" and "7" can be used for the pixel whose area data is "1". That is, the ZONE signal “0” is the thinnest character, ZON.
Although the E signal "7" represents a halftone / halftone image, this binary processing may be performed.

The area data is not limited to being specified by the digitizer as described above. For example, by providing an interface with an external device so that image data can be input from an external device such as an external storage device, Area data from the device may be used.

In the above embodiment, as shown in FIG. 6, the RGB signal is used as the input to the black character determining unit 113, but the present invention is not limited to this. For example, the CMY signal output from the LOG converting unit 107 is used. You may use it.

Further, in the above embodiment, the black character determination unit 1
The input to the character thickness determination circuit 114 constituting 13 is
RGB signals are used. However, not limited to this,
For example, as shown in FIG. 27, an L signal may be obtained through the Lab converter 2010, and the L signal may be used to perform subsequent processing. 26, the same reference numerals are used for the same constituent elements of the character thickness determination circuit shown in FIG.

The present invention may be applied to a system composed of a plurality of devices or an apparatus composed of a single device. Further, it goes without saying that the present invention can also be applied when it is achieved by supplying a program to a system or an apparatus.

[0125]

As described above, according to the present invention,
Judgment parameter and processing for black character processing by adjusting the character / photo separation level when determining the thickness of the character / line drawing part in the image and performing image processing by combining outline information and saturation information of the character / line drawing. By simultaneously changing the degree of, the thickness of the character to be processed and the degree of processing change, and it becomes possible to reproduce the image according to the user's preference.

Further, black character processing can be arbitrarily performed for each area.

[Brief description of drawings]

FIG. 1 is a diagram showing a cross-sectional configuration of an image processing apparatus according to an embodiment of the present invention.

FIG. 2 is a diagram showing an external configuration of a CCD 210.

3 is a cross-sectional view when the image scanner unit 201 is cut along a dotted line aa 'shown in FIG.

FIG. 4 is an enlarged view of the light receiving element indicated by reference numeral B in FIG.

FIG. 5 is a timing chart showing a control operation of density reproduction in the printer unit according to the embodiment.

FIG. 6 is a block diagram showing the flow of image signals in the image signal processing unit 209 of the image scanner unit 201 according to the embodiment.

7 is a diagram showing the timing of each control signal in the image signal processing unit 209 shown in FIG.

FIG. 8 is a block diagram showing an internal configuration of an edge detection circuit 115.

FIG. 9 is a diagram for explaining a character thickness determination circuit 114.

FIG. 10 is a diagram showing a state of line delay due to a FIFO and a Laplacian filter.

FIG. 11 is a diagram showing an example of edge detection.

12 is a block diagram showing a detailed configuration of a saturation determination circuit 116 that constitutes a black character determination unit 113. FIG.

FIG. 13 is a diagram showing data conversion characteristics in an LUT.

FIG. 14 is a block diagram showing a configuration of a character thickness determination circuit 114 that constitutes a black character determination unit 113.

FIG. 15 is a block diagram showing an internal configuration of a character / halftone detection circuit 2013.

16 is a block diagram showing a detailed configuration of a halftone dot area detection unit 2014. FIG.

FIG. 17 is a diagram showing a rule of edge direction detection in the edge direction detection circuit 2044.

FIG. 18 is a diagram showing a rule of opposite edge detection.

FIG. 19 is a diagram showing a shape of a window in a counting unit 2049.

20 is a block diagram showing an internal configuration of an area size determination circuit 2016. FIG.

FIG. 21 is a block diagram showing a configuration of a ZONE correction unit 2084.

FIG. 22 is a diagram showing how the output of PICT_FH is determined according to the thickness of a character.

FIG. 23 is a diagram showing an encoding rule of an encoder 2083.

FIG. 24 is a diagram showing an algorithm for character detection in halftone dots / halftones.

FIG. 25 is a diagram specifically showing a state of processing by the algorithm shown in FIG. 23.

FIG. 26 is a diagram showing the contents of input / output correspondence of the LUT 117.

FIG. 27 is a block diagram showing a modified example of the character thickness determination circuit 114.

FIG. 28 is a diagram showing a masking UCR arithmetic expression.

FIG. 29 is a diagram showing a display for adjusting a character / photo separation level.

FIG. 30 is a diagram showing determination parameters.

FIG. 31 is a diagram showing an LUT showing the degree of character processing.

FIG. 32 is a diagram showing an LUT showing the degree of character processing.

FIG. 33 is a diagram showing an LUT showing the degree of character processing.

FIG. 34 is a diagram showing an LUT showing the degree of character processing.

FIG. 35 is a diagram showing an LUT showing the degree of character processing.

FIG. 36 is a diagram showing an LUT showing the degree of character processing.

FIG. 37 is a diagram showing an LUT showing the degree of character processing.

FIG. 38 is a diagram showing an LUT showing the degree of character processing.

FIG. 39 is a diagram showing a display of a document mode on the operation unit.

FIG. 40 is a diagram illustrating a configuration of a second example of the present application.

Claims (10)

[Claims] 1. An image processing apparatus for processing image data representing a document image, a determining unit for determining a character / line drawing in the document image from the image data, and a predetermined unit based on a determination result of the determining unit. Image processing means, a designation means for designating in a multi-step manner which one of a character / line drawing and halftone should be emphasized when extracting features of a document image, and the determination based on the designation by the designation means. An image processing apparatus comprising: a parameter for character / line drawing determination by means, and control means for controlling processing by the processing means. 2. The image processing apparatus, wherein the determining means determines the thickness of a character / line drawing based on a pixel of interest and a pixel adjacent to the pixel of interest. 3. The image processing apparatus according to claim 1, wherein the image data is full-color image data. 4. The image processing apparatus according to claim 1, wherein the original image is an optically read image. 5. The original image has four color components of yellow, cyan, magenta, and black, and the processing means prepares a plurality of undercolors for signal conversion of the four colors. The image processing apparatus according to claim 1, wherein the proportion of the removal amount is changed according to the thickness of the character / line drawing. 6. The method according to claim 1, wherein the processing unit changes a plurality of types of region widths of a reference region for determining a contour prepared in advance, depending on the thickness of a character / line drawing. The image processing device described. 7. The image processing apparatus according to claim 1, wherein the processing unit switches a plurality of spatial filter coefficients prepared in advance according to the thickness of a character / line drawing. 8. The processing means, for a character / line drawing,
The image processing apparatus according to claim 1, wherein different image processing is performed depending on whether the character / line drawing is in a white background or in a halftone or halftone. 9. An image processing apparatus for processing color image data representing a document image, detecting means for detecting a line image portion of a specific color from the color image data, and signal generating means for generating predetermined area data. An image processing apparatus comprising: a control unit that controls a parameter for detection by the detection unit according to the area data. 10. The image processing apparatus according to claim 9, wherein the signal generating unit includes a digitizer.