JPH0364257A - Image processor - Google Patents

Abstract

PURPOSE:To improve the discriminating capacity of a character part and a half tone image part by controlling an operation of a means for discriminating a half tone part with regard to a designated area in an object image. CONSTITUTION:The subject processor is provided with a feature extracting part 403 constituted of a color deciding part 106 and a character edge deciding part 107, a control part 401, and an operating part 407. In this state, with regard to a designated area in an object area, a discriminating condition of a discriminating means for discriminating a character part of a half tone part is set. In such a way, as for an area having possibility that an erroneous discrimination is executed at the time of discrimination, a satisfactory discrimination can be executed by controlling the discriminating means, for instance, inputting in advance a discriminating condition conforming to such an area.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to an image processing apparatus, and particularly to an image processing apparatus capable of performing image area determination.

[Conventional technology]

Conventionally, the text parts and halftone image parts in the target image are distinguished, the text parts are subjected to simple binarization, and the halftone parts are treated with the characteristics of each image by applying, for example, a systematic dithering method. It was constructed so that good images could be obtained.

[Problem to be solved by the invention]

However, even in the conventional method, there is still much room for improvement in terms of the ability to discriminate between character areas and halftone image areas. In view of this point, it is an object of the present invention to improve the discrimination ability of the literary halftone image area regarding the text portion, the halftone portion, that is, the specific area in the image.

[Means to solve the problem]

The present invention has been made in view of the above-mentioned points, and includes a determining means for determining a character part or a halftone part in a target image, a specifying means for specifying an image area in the target image, and a specifying means for specifying an image area in the target image. and control means for controlling the operation of the discriminating means with respect to the determined area.

[Effect]

In the above configuration, the determination means is controlled by the control means with respect to the area designated by the designation means.

(Margin below) (Example) The present invention will be described below using a full-color digital copying machine as an example. However, the present invention is not limited to such an example, and the present invention can be applied to various devices such as a device having only the function of converting a target image into an electrical signal. It is also applicable to

〔overall structure〕

FIG. 2 shows an overall configuration diagram of a full-color digital copying machine.

Reference numeral 201 denotes an image scanner section that reads a document and performs digital signal processing. Further, 202 is a printer unit, which prints out an image corresponding to the original image read by the image scanner unit 201 on paper in full color.

In the image scanner unit 201, 200 is a mirror pressure plate, and an original 204 on an original platen glass (hereinafter referred to as platen) 203 is irradiated with a lamp 205, and mirrors 206, 2
07. 208, an image is formed on a 3-line sensor (hereinafter referred to as C0D) 210 by a lens 209, and full color information red (R), green (G), blue (B) l
It is sent to the signal processing unit 211 as ff1 minute. In addition, 20
5, 206 is the velocity V, 207.208 is l / 2
The entire surface of the document is scanned by mechanically moving in a direction perpendicular to the electrical scanning direction of the line sensor at a speed of .

The signal processing unit 211 electrically processes the read signals to produce magenta (M), cyan (C), and yellow (Y) signals.

The printer unit 202 decomposes black (Bk) into each component.
send to Furthermore, for each document scan by the image scanner section 201, one of M, C, Y, and Bk components is sent to the printer section 202, and one printout is completed by scanning the document four times in total.

M sent from the image scanner unit 201.

The C, Y or Bk image signal is sent to the laser driver 212. The laser driver 212 modulates and drives the semiconductor laser 213 according to the image signal. The laser beam scans the photosensitive drum 217 via a polygon mirror 214, an f-theta lens 215, and a meter 216.

Reference numeral 218 denotes a rotary developing unit, which is composed of a magenta developing unit 219, a cyan developing unit 220, a yellow developing unit 221, and a black developing unit 222. Develop the electrostatic latent image with toner.

223 is a transfer drum, and paper cassette 224 or 225
The paper fed from the transfer drum 223 is wound around the transfer drum 223, and the image developed on the photosensitive drum 217 is transferred onto the paper.

After the four colors M, C, Y, and Bk are sequentially transferred in this manner, the paper passes through the fixing unit 226 and is discharged.

[Image scanner]

FIG. 3 is an internal block diagram of the image scanner section.

In FIG. 3, 101 is a counter, and CCD 21
A main scanning address 102 specifying the main scanning position of 0 is output. That is, when the horizontal synchronization signal H3YNC is 1, it is set to a predetermined value by a CPU (not shown), and is incremented by a pixel clock signal CLK.

The image formed on the CCD 201 is charged and converted by three line sensors 301, 302, and 303, and read signals for Rfi, G, and Bd are sent to amplifiers 304, 305, and 306, and a sample hold circuit. 30
7.308, 309 and A/D converter 310. 311
8-bit digital image signal 313 for each color through ゜312
(R), 314 (G), and 315 (B).

[Signal flow]

Figure 4 shows the overall signal flow. Components common to those in FIG. 2 are designated by the same numbers. In the figure, CLK is a clock signal for transferring pixels, H5YNC is a horizontal synchronization signal, which is a synchronization signal for starting main scanning, and CLK4 is a clock signal for generating a 400-line screen, which will be described later, as shown in FIG. The control unit 401 controls the image scanner unit 2011, the signal processing unit 211, and the printer unit 20.
Sent to 2.

An image scanner unit 201 reads a document 204 and converts R, G, and B signals as electrical signals to a color signal processing unit 402.
and sends it to the feature extraction unit 403. The feature extracting unit 403 sends a COL signal to the color processing control signal generating unit 404 indicating that the currently processed pixel is a black image, and is a BL double signal indicating that the image has a tint. A UNK signal indicating that there is a possibility that the image is a colored image, a CAN signal for canceling the BL double signal, and an EDGE signal indicating a character edge are sent.

The ATLAS signal output from the control unit 401 is an image processing operation switching signal when copying a document with fine text such as a map, and is input to the feature extraction unit 403 and the color signal processing unit 402.

Similarly, the 4-bit SEG output from the control unit 401
The signal is a control signal that changes the degree of character extraction, and is input to the feature extraction unit 403.

Reference numeral 407 denotes an operation panel through which the CPU inputs manual input of keys to the control unit 401 and controls display operations.

FIG. 6 shows details of the operation panel 407. In FIG. 6, 601 is a dot matrix liquid crystal display section with 64×192 dots. 602 is a copy start key, 603 is a recording paper cassette selection key, 604 is a part of the number pad, 60
5 is a clear key and a copy operation stop key manually operated by the numeric keypad. 606 is a key to reset the set display, 607 to 610 are keys for moving the cursor on the liquid crystal display up,
A key 611 for moving in the downward, left, and right directions is a key for ending selection on the liquid crystal display section. 612 is an asterisk (*) key for setting various copy modes, and 613 is an image key for setting image editing mode.
It is the creation key.

Returning to FIG. 4 again, the color processing control signal generation section 404 receives the above signal from the feature extraction section 403 and generates a color processing control signal for the color signal processing section. These are two multiplication coefficient signals GA for calculating weighting of two types of image signals.
FIL to switch INI, GAIN2 and spatial filter
This is a CAM signal that switches between a signal and a plurality of density conversion characteristics. The control unit 401 sends 2b to each processing block.
The PHASE signal of it is sent. This signal corresponds to the developed color of the printer section, and is 0.1 to 0.1 of the PHASE signal.
2.3 are the developing colors magenta (M), cyan (C),
It means yellow (Y) and black (Bk).

The color signal processing section generates a recording image signal V to the printer section 202 based on this PHASE signal and the color processing control signal.
Generates IDEO.

Based on this VIDEO signal, the printer unit 202 pulse-width-modulates the laser emission time and outputs a copy output 406 with gradation.

An SCR signal GA is input to the printer unit 202 from a color processing control signal generation unit 404 . The printer unit 202 is
Using this SCR signal, a plurality of pulse width modulation basic clocks (screen clocks) are switched to express the optimum density for the original. In this embodiment, when the SCR signal is 0, pulse width modulation is performed in units of 1 pixel, and when the SCR signal is 1, pulse width modulation is performed in units of 2 pixels.

The operations of the color signal processing section 402, feature extraction section 403, and color processing control signal generation section 404 will be described in detail below with reference to FIG.

[Feature extraction part]

The feature extraction unit 403 is composed of a color determination unit 106 and a character edge determination unit 107.

FIG. 11 shows the configuration of each processing section.

In FIG. 11, 1101 is a pixel color column section, and each pixel has a BLP signal indicating that it is black, a C0LP signal indicating that it has a tint, and a UNK P signal indicating that it is unknown which one it is. Generates a signal and area processing unit 1
Send to 102. The region processing unit 1102 determines the BLP, C0LP, UNKP, and G signals for each region within a 5×5 area and removes errors from the BL.

Generates COL, UNK, and CAMA.

Reference numeral 1103 denotes a character edge determination unit which determines whether or not it is a character edge portion based on the G signal and generates an EDGEO signal. The reason for determining whether or not it is a character edge portion using only the G signal is that, as shown in Fig. 12, the G signal is closest to the visibility characteristics among the R, G, and H signals. This is because the signal can be represented by a character edge detection signal of a white/black image.

1104 is a halftone dot determination unit, and a character edge determination unit 110
Based on the density direction signal DSL from 3, a DOT signal is output which determines, pixel by pixel, that the pixel of interest is included in the halftone dot area. When the document is a halftone dot printed matter, the character edge determination unit 1103 often determines the halftone dots as characters. In this embodiment, processing is performed on character edges, such as edge emphasis and increasing recording resolution, in order to improve the sharpness of the recorded image, as will be described later. If such processing is applied to a halftone image, moiré will occur and the quality of the recorded image will be significantly degraded. Therefore, it is determined by this halftone dot determination signal DOT that the document has a halftone dot area, and the gate 1105 prevents the generation of the character edge signal EDGE.

The ATLAS signal and the SEG signal are output from the control section 407. As will be explained in detail later, the ATLAS signal is a control signal for recording fine characters clearly.
The SEG signal is a signal that variably controls the slice level for character edge detection.

FIG. 13 is a block diagram of saturation determination by the pixel color determination unit 1101.

In FIG. 13, numeral 301 is a MAX/MIN detector, 1302 to 1309 are selectors, and 1310 to 13
15 is a subtracter which outputs A-B for input A and input B. 1316 to 1323 are comparators, and 1316 for input A and input B. 1319 is 13 when 2A>H
17°1320.1322.1323 is AGE,
1318 and 1321 output 1 when A>2B, and output 0 otherwise. 1324-1328 are AND gates, 1329 is NOR gate, 1330 is N
ANDA-to.

In the above configuration, the circuit shown in FIG. 14-1 is used for the MAX/MIN detector 1301. In FIG. 14-1, 1350, 1351, and 1352 are comparators that output 1 when R>G, G>B, and BAR, respectively. The circuit shown in FIG. 14-1 uses the following determination signals SOO and SOI, as shown in FIG. 14-2.

Generate SO2, SIO, Sll, and S12. That is, when MAX is R or when R, G, and B are all equal, 5OO=l, 5O1=SO2=O.

If MAX is G, 5ot=t; 5OO=SO2=
If 0°MAX is B, 5O2=l, 5OO=SO
1=O.

If MIN is R, or if R, G, and B are all equal, 510=1,511=S12=0° MIN is G
In this case, 511=1.510=SI2=O.

If MIN is B10), 512=1.510=S11
=O.

becomes.

For example, when MAX is R, comparator 1350 outputs l and comparator 1352 outputs O because R>G or R2H. ANDI outputs 1, and ORI outputs 1. AND2. AND3 outputs O. That is, 5OO=1, 5O1=SO2=0. The results of similar determinations are shown in the table shown in FIG. 14-2.

MAX/MIN detector output SOO, SQL, SO2
is input to the selector 1302, and the outputs SIO, Sll,
512 is input to selectors 1303-1309.

The selectors 1302 to 1309 are composed of an AND circuit and an OR circuit as shown in FIG. 15-1. According to this selector, inputs A and B, as shown in Figure 15-2.

For C, output A when 5O=1, 5L=32=O, output B when 51=1, 5O=S2=O, and output 52
Outputs C when =1,5O=S1=O. In this embodiment, inputs A, B, and C correspond to R, G, and B signals.

In the pixel color determination of this embodiment, the maximum value of the R, G, and B signals is set as MAX, and the minimum value is set as MIN, and the values of A, B, C, and D are determined as shown in Figure 16-1. This is done by dividing it into four areas.

In other words, in the achromatic color area, the smaller the difference between MAX and MIN is, and the closer it is to a chromatic color, the more M A,
MAX by taking advantage of the fact that the difference between X and MIN increases.

M by linear simultaneous inequalities with MIN as a parameter
Divide the AX-MIN plane.

Specifically, ka, kb, kc, ia, ib,
ic, WMX.

Let WMN be a predetermined constant and divide it into four areas A, B, C, and D as shown in Figure 16-1.

A is a dark achromatic (black) area. (MAX.

The condition for MIN) to be included in this area is MIN≦WMN
The condition that MAX≦WMX and MIN) is included in this area is MIN≦WMN
or MAX≦WMX, and either and all of the above are satisfied.

C is a chromatic color area. The conditions for (MAX, MIN) to be included in this region are that MIN≦WMN or MAX≦WMX, and all of the following are satisfied.

B is an area between dark achromatic colors and chromatic colors. (MAX
.

It is to satisfy one of the following.

D is a bright achromatic (white) area. (MAX.

The conditions for MIN) to be included in this area are to satisfy both of the following.

Figure 16-2 shows the output signals for each state of A, B, (1:, D). That is, A
If included in the area, BLP=I, UNKP=COLP=O1B If included in the area, UNKP=1. BLP=COLP=0, if included in the C area, C0LP=1. BLP=UNKP=0, and when included in area D, BLP=1, UNKP=COLP=0.

The areas 1304 to 133 in FIG. 13 perform the above area determination.
0 circuit. Depending on the output of the MAX/MIN detector 1301, selectors 1302 and 1303 select the MAX signal and MIN signal from among R, G, and B, respectively.
Selectors 1304 to 130 are linked to selector 1303.
9 are also constants ka, kb, kc, ia, respectively.
Select values for ib and ic. For example, MAX is an R signal,
When MIN is a G signal, the selector 1304 is KAG,
1305 is KBG.

1306 selects KCG, 1307 selects iAG, 1308 selects iBG, and 1309 selects iCG, each with constant k.
a, kb, kc.

Let ia, ib, ic. In this way, the minimum value is R
,G.

Constants ka, kb, kc, i depending on one of B
a, ib.

The reason for changing the value of ic is as follows.

Now, the color space separation in No. 16-l uses the R, G, and B color separation signals of the CCD sensor. This R,G
, the MAX and MIN planes of the B signal have deviations from the human visual sensitivity characteristics. That is, it is necessary to switch the delineation between the achromatic color area and the chromatic color area depending on the color of the document.

Therefore, in this embodiment, ka and kb are determined depending on the original color.

The MAX axis intercept values of kc, ia, ib, and ic are variable. In order to specify the original color, this embodiment uses the determination result of which of the R, G, and H light quantity signals is the MIN signal. This is due to the following reasons. The color tone of the original judged by humans is C, M included in the original.
, Y, and the maximum color of the reflection density corresponds to the minimum color of the light amount signal. Also, when converting R, G, and B light amount signals to C9M and Y density signals, the -10g function is used, so the range is compressed on the maximum value side of the light amount signal, and the range is expanded on the minimum value side of the light amount signal. . In this way, it is advantageous in terms of determination accuracy to use the MIN color signal of the light amount signal to separate the color signal that governs the color tone in the density signal.

Therefore, Figure 15-1 shows the details of the configuration of selector 1.
In 304 to 1309, decode signals AIO, Sll, and S12 indicating the MIN color are used to determine M according to the MIN color.
AX intercept value ka, kb, kc, ia, ib,
Generate ic.

In this example, ka, kb, and kc are determined by experimentally determined values taking into account the color separation filter of the CCD sensor.

The following values are set for ia, ib, and ic. however,
The range of RIG and B is from 0 to 255.

As described above, the subtracters 1316 to 1315 subtract from the MAX value using different MAX axis intercept values for each MIN color. In the comparator 1316, 2XMrN>(M A
X-k a ) is determined and it is detected that the combination of the MAX value and the MIN value is above the straight line S in FIG. 16-1. Similarly, the comparators 1317 to 1321 each have a set of MAX value and MIN value on straight lines t and u, respectively.

v, w, x is detected.

Also, the MAX value at comparator 1322.1323,
It is detected that the MIN value is larger than the predetermined values WMX and WMN, and a gate 1324 performs AND processing to generate a WB multiplied signal indicating that the read pixel is a white skin portion.

By encoding the above signal as follows, BLI
, UNKI, and C0LI signals are generated. Since the BE,1 signal is in the area A in Figure 16-1, the AND gate 1325 detects that it is above the straight lines s, t, and u, and
Game) 1326 adds the condition that it is not in the D area.

NA indicates that the C0LI signal is below the straight lines v, w, x.
AND the condition detected by the ND gate 133o and not in the D region
It is added at gate 1328.

The UNKI signal is below the straight lines S, t, and u and the straight line V
.

A NOR gate 1329 detects that it is above W and X, and an AND gate 1327 adds a condition that it is not in the D region.

[Area processing section]

FIG. 7 shows a block diagram of the area processing unit 11o2 shown in FIG. 11.

BLP determined by the pixel color determination unit 1101.

The C0LP and UNKP signals are sent to the line memories 1701, 1702, 1703. The lines are delayed by 17041m and synchronized by the H3YNC signal and cLK signal shown in Figure 3, and 5 lines are output simultaneously. here,
BKP.

C0LP and UNKP delayed by one line are respectively BE2. C0L2. UNK2. BE3. C0L3. UNK3. BE4. C0L4. UNK4. BE5. C0L5. When UNK5 is used, each signal is delayed by 5 pixels at 1705. 8th
The number of black pixels (BL) is counted within the 5×5 area shown in the figure to obtain NB, and similarly, the number of chromatic pixels (COL) is counted by counting means 1706 to obtain NC.

Further, a comparator 1707 compares the number NB of black pixels and the number NC of chromatic pixels within the 5×5 block.

Furthermore, gate circuit 1708. 1709. 1710.
1711゜1712.1713, 1714.1715
The output of the pixel color determination unit for the center pixel of the 5×5 area is BK3°C0L3. Calculated together with the result of UNK3, a BL multiplied signal indicating that the center pixel is black, a COL signal indicating that the center pixel is chromatic, and a UNK signal indicating that the center pixel has intermediate saturation are output. . The determination criterion at this time does not overturn the determination for those whose determination result according to the first determination criterion is a black pixel or a chromatic pixel. That is, when BL3=1 or C0L3=1, BL=1 or C0L=1. Furthermore, if the judgment result of the first judgment criterion is between chromatic pixels and achromatic pixels, a comparator 1716 judges whether the number of black pixels is greater than or equal to a predetermined value (NEC), Comparator 171
In step 7, it is determined whether the number of chromatic pixels is greater than or equal to a predetermined value. Furthermore, a comparator 1707 determines whether the number of black pixels or the number of chromatic pixels is greater. If the number of black pixels is greater than a predetermined value and NB>NC, that is, the pixel of interest is UNK.
Even so, if there are many black pixels in the 5×5 matrix including the pixel of interest, UNK3 becomes BL in gate F08.

In addition, if the number of chromatic pixels is more than a predetermined value and NB≦NC, that is, even if the pixel of interest is UNK, the number of chromatic pixels including the pixel of interest is 5
If there are many chromatic pixels in the ×5 matrix, gate 17
At 09, UNK3 becomes COL.

In this embodiment, the scanning optical system 206, 207, 208
In order to remove color blurring at color change points of the original due to scanning unevenness and magnification error of the imaging optical system 209, chromatic and achromatic colors are determined by the algorithm described above. If there are no more than a predetermined number of black pixels or chromatic pixels around the UNK3 signal, the gate 1713. 1
714. 1715 and outputs an intermediate saturation signal UNK.

Next, FIG. 17-1 shows the configuration of a CAN signal generating section included in the area processing section shown in FIG. 11.

In the logic circuit for generating the BL multiplied signal shown in FIG. 7, if the pixel of interest is a black pixel, the BL multiplied signal is output regardless of its surroundings. However, if there is the aforementioned scanning speed unevenness or imaging magnification error, a black signal (Bk) may occur around the color signal (C) due to color fringing, as shown in FIG. 9. A black signal (Bk) due to this color fringing (C) is generated around the color signal as shown in FIG. 10, so the light amount value is larger than that of the color signal. Therefore, the CAN signal generation section shown in FIG. 7-1 detects whether a color signal (COL) having a smaller light amount value than the pixel of interest exists around the pixel of interest and generates a CAM signal.

In this embodiment, the G signal closest to the above-mentioned visibility characteristic is used as the light amount signal. This G signal is converted into one line of fi.
o Memories 1718, 1719, and 1720 delay the target line G3 signal and G2.
The G4 signal is input to the calculation section 1722.

At the same time, the 3-line color determination signal C0 created in Figure 7
L2. C0L3. Enter C0L4.

Details of the calculation unit 1722 are shown in FIG. 17-2.

G2. G3. G4. C0L2. C0L3. C
0L4 is delayed by two or three pixels by flip-flops shown at 1723 to 1735, respectively. Here, the pixels of interest are G32 and C0L32. G32 is selected by the comparators 1737 to 1740 as the peripheral pixel G22.
G31. G33. It is compared with G42. The comparator outputs H when the light amount value of the surrounding pixels is lower than that of the pixel of interest. And at AND gates 1741-1744,
AND the color judgment signal of the surrounding pixels and OR gate 1
A CAN signal is output at 745.

That is, if the level around the pixel of interest is lower than the level of the pixel of interest and there is a brown component, it is determined that color fringing as shown in FIGS. 9 and 10 has been derived, and a CAN signal is generated.

This is convenient in order to prevent the occurrence of "black bleeding" around colored characters, which occurs when processing electrical signals obtained by reading characters such as "maroon", for example.

[Character edge determination section]

Next, the operation of the character edge determination section will be explained using FIG. 19.

The manuscript 1901 shown in (a) in FIG. 19 showing a conceptual diagram is
This is an example of an image with shading, and a character edge area 1902
and a halftone area 1903 expressed by halftone dots. As a method for extracting edge information in an image, in this embodiment, as shown in 1904, it is determined whether there is a sharp density change in a pixel block whose unit is nine neighboring pixels surrounding the pixel of interest X i, j. A determination is made as to whether or not, and further,
It takes advantage of the fact that steep concentration change points exist continuously in a specific direction.

Specifically, the pixel of interest X11. , take the difference value of the neighboring pixel, take the parameter expressed by Determine whether they exist continuously in a specific direction. Note that X iJ, etc. are the pixel of interest and peripheral pixels as shown in (b) in FIG.

Specifically, the detection of a vertical edge with a high concentration on the right side as shown at 1905 in FIG. Yes (Figure 21 2101 and 2102). Detection of a horizontal edge with a high concentration on the lower side as shown in 1906 has a property that points having a large value of J2 in equation 6 are continuous in the horizontal direction (FIG. 21, 2103 and 2104).

Detection of an edge in the diagonal right direction with high concentration in the lower right direction, as shown in 1907, has the property that the points where the value of 13 in equation 6 is large are continuous in the diagonal right direction (Fig. 21, 2105. 2106). Detection of an edge in the left diagonal direction with high concentration in the lower left direction as shown in 1908 has a property that the points where the value of 14 in the sixth equation is large are continuous in the left diagonal direction (Fig. 21 2107° 2108 ). 190
Detection of a vertical edge with a high concentration on the left side as shown in FIG. 9 has a property that points having a large value of J5 in equation 6 are continuous in the vertical direction (FIG. 21, 2109 and 2110).

Detection of a horizontal edge with high concentration on the upper side as shown in 1910 has a property that points having a large value of 16 in equation 6 are continuous in the horizontal direction (Fig. 21 2111.211
2). Detection of an edge in the right diagonal direction with high concentration in the upper left direction as shown in 1911 has a property that the points where the value of J7 in the sixth equation is large are continuous in the right diagonal direction (Fig. 21 2113, 2114 ). Detection of an edge in the left diagonal direction with high concentration in the upper right direction as shown in 1912 has a property that the points where the value of J8 in the sixth equation is large are continuous in the left diagonal direction (Fig. 21 2115, 2116 ).

On the other hand, even in the halftone dots shown in 1909 to 1912, the values from 11 to J4 are thick (become thicker).Furthermore, as the size of the halftone dots increases, continuity in a specific direction also occurs, so it is incorrectly judged as a character edge. You will end up being rejected.

This halftone image has density symmetry as shown in FIG. 22-2 (details will be described later). In this embodiment, the means for extracting the features of the halftone dot image is configured to cancel the detection result of a character edge when it is determined that the halftone dot image is a halftone dot.

FIG. 18-1 shows a block diagram of the character edge determination section.

In FIG. 18-1, 1801 is a density change detection unit, and 1802 is a part that detects continuity of density changes for extracting character edges. Reference numeral 1842 denotes a halftone determination unit that detects that the pixel of interest is a halftone image, which includes a halftone feature extraction unit 1827 and a halftone area determination unit 1828 (the internal details will be described later). When the halftone detection signal DOT becomes "L", the output of the NAND gate 1840 becomes 0 (in the case of ATLAS20), and the AND gate 1
．． 841, the character edge determination signal EDGEO is canceled and EDGE=0. That is, even if a portion is determined to have an edge, if it is determined to be a halftone dot, it is excluded from the edge, and the character edge determination signal becomes "0".

However, in manuscripts such as maps, minute characters are written in the halftone image. Therefore, for example, if the map mode is selected by the operator via the operation unit 407 and ATLAS=1 is set by the control unit 401 in response, N
The DOT signal is canceled by the AND gate 1840, and the character edge information in the halftone dot is output as EDGE=1.

Next, the concentration change point detection unit 1801 shown in FIG. 18-1 will be explained below.

The image signal G is converted into a TXG signal by a signal conversion table 1826. The configuration of the signal conversion table 1826 is
Shown in Figure 8-8.

In Figure 18-8, the signal conversion table is 1881 and 1.
There are two types of 882. The relationship between the input and output of table 1881 is configured as shown in the following equation.

Here, both the input and output of the table are 8-bit signals, and the signal values range from 0 to 255. This table 18
Reference numeral 81 is used to determine the edge of a character in a document containing a mixture of various types of information such as a normal color photograph and a halftone photograph of one character.

In the first place, the character edge determination section normally separates character information in a document from other photographic information. Normal character information is often recorded on a white background. On the other hand, photographic information is recorded as continuous changes in density information, and there are almost no sharp changes in density in a white background.

Therefore, in table 1881, in order to make the character information in the white background easier to understand, the level change of the information near the white background (level 255) is set to be large. In order to make it difficult to detect density changes in the background with a certain level of density, which are often seen in photographs, as character edges, level changes in information near the black background (level 0) are compressed. Therefore table 1
In 881, the range of O≦X≦1 is used for the y-X2 characteristic.

On the other hand, table 1882 does not perform input and output level conversion as shown in the following equation.

out = in This signal conversion table 1882 is for separating minute character information recorded during the map color navigation. Therefore, input = output so that character information on a colored background and character information on a white background are equally separated.

The outputs of these two conversion tables are selected by selector 1883 and become TXG signals. The ATLAS signal is input as a selection signal to the selector 1883, and the output of 1882 is selected at H level, and the output of 1881 is selected at L level.

T output from selector 1883 shown in Figure 18-8
The XG signal is delayed by line memories 1803 and 1804, and three lines are simultaneously output to the detector 18 shown in Fig. 18-1.
05 (the inside is shown in FIG. 20-1), and eight types of density change information AKI-AK8 are output. Here, they are expressed as (below the margin), respectively, to the right, bottom right, and
When there is a steep increase in density in the lower left, left, upper, upper left, and upper right directions, it is l, and otherwise it is 0.

Here, T1 is the main scanning direction density change detection slice level;
T2 is a slice level for detecting a density change in the sub-scanning direction, T3 is a slice level for detecting a density change in a diagonal direction, and ATL
It is variably controlled by the US signal and the 4-bit SEG signal. Note that the SEG signal is data input by the user from the operation unit 407 shown in the first diagram.

As shown in FIG. 20-1, the detector 1805 includes flip-flops 2001 to 2006 and difference calculators 2007 to 2o.
14, comparators 2015 to 2122.

That is, in FIG. 20-1, the flip-flops 2001~
9 (2006) by the image clock CLK.
The image data of the pixel shown in 1904 in b) is latched,
In the difference calculators 2007 to 2014, the above J1 to
J8 is calculated, and the comparators 2015 to 2022 output determination results AKI to AK8. 2023 is a density change detection slice level generation unit which inputs the ATLUS signal and the SEG signal as addresses, and outputs T1T2 . T
This is a ROM table that outputs 3 as data. The contents of this table are shown in Figure 20-2.

In this embodiment, the SEG signal changes in nine stages from 0 to 8. As this value increases, the slice level T l
r T2+ T3 also increases. As a result, the density change signals AKI to AK8 will not be generated unless there is a large density change in the document. Conversely, if the SEG signal value is decreased, T
In 72+ "T3 becomes small and AKI-AK8 occurs due to a small density change in the document.

That is, in this embodiment, the degree of detection of density changes is changed by controlling the SEG signal. Also ATL
When the AS signal is 1, compared to when the ATLAS signal is 0, T, , T2 . The T3 value is approximately halved overall, making it easier to detect minute density changes in the document.

As a result, when the ATLAS signal is 1, minute character information on the document is also detected.

Returning to FIG. 18-1, 1802 is a part that determines whether a steep density change is continuous in a direction having an angle of 90° with respect to the direction of the density change. In this example, as shown in Figure 22-1, 5x5 pixels centered on the pixel of interest are
We are looking at a series of density changes within a pixel block. For example, 2201 shown in FIG. 22-1.

Reference numeral 2202 indicates a reference pixel when detecting a succession of edges in the vertical direction. The above-mentioned characteristics of density changes of peripheral pixels are AK
This is a reference pixel when detecting three consecutive pixels that are I or AK5. 2203゜2204
Similarly, is a reference pixel when detecting consecutive AK2 or AK6. 2205.2206 is for detecting consecutive AK4 or AK8, and 2207°22
08 is a reference pixel when detecting consecutive AK3 or AK4.

In this embodiment, the reason why the pixel of interest is not brought to the center of the continuity check when extracting a succession of density changes is as follows. That is, as shown in FIG. 22-3, pixels forming the end of a character are also determined as pixels included in a continuous edge.

In order to detect a succession of density changes in the above-mentioned 5.times.5 area, edges in eight directions for each pixel detected by the density change point detection section are delayed by four lines of line memories 1805 to 1808. Density change information for 5 lines formed in this way AKI to AK8, BKI to BK8, CK
I to CK8, DKI to DK8, and EKI to EK8 are each delayed by pixels in stages 1 to 5 of flip-flops in order to check the continuity shown in FIG. 22-1. After that, the center pixel (CP
U3. CBr4. CLF3. CRT3゜CUL3. C.B.
A continuation of three pixels having edges (R3, CUR3, CBL3) is detected, and a NOR gate 1825 generates an EDGEO signal indicating that the center pixel constitutes a continuous edge. For example, gate 1809 has the 22nd-1st feature of AK6.
Continuity is detected in the form 2203 shown in the figure.

In addition, the 1810 has the characteristics of AK6 as shown in Figure 22-1.
It has been detected that the numbers are continuous in the form of 4. Similarly, gate 1811 detects that the features of AK2 are continuous in the form 2203. Gate 1812 detects that the features of AK2 are continuous in the form 2204. Gate 18]3 detects that the features of AK5 are continuous in the form of 2201. Gate 1814 detects that the features of AK5 are continuous in the form 2202.

Gate 1815 detects that the AKI features are continuous in the form 2201. Gate 1816 detects that the AKI features are continuous in the form 2202. Gate 1817 detects that the features of AK7 are continuous in the form 2208.

Gate 1818 detects that the features of AK7 are continuous in the form of 2207. Gate 1819 detects that the features of AK3 are continuous in the form 2208. Gate 1820 detects that the features of AK3 are continuous in the form 2207. Gate 1821 is AK8
It is detected that the features are continuous in the form 2205. Gate 1822 detects that the features of AK8 are continuous in the form of 2206. Gate 1823 is AK
It is detected that the features of 4 are consecutive in the form 2205. Gate 1824 detects that the features of AK4 are continuous in the form 2206.

In this way, only the character edge part in the character area 1902 shown in FIG. 19 is determined as an EDGE signal and output.

[Halfpoint determination section]

The halftone dot determination section 1842 shown in FIG. 18-1 is the 23-2
As shown in the figure, a halftone feature extraction unit 1827 detects symmetric density changes of halftone dots, and a halftone dot area detects that the halftone feature signal DOTO is distributed over a certain number within an area of a certain size. It is composed of a detection section 1828.

This is because even in complex character information such as kanji, there are parts that exhibit symmetrical density changes as shown in Figure 23-2. Symmetrical density changes do not exist in a wide range in this complex character, but symmetrical density changes are distributed over a wide range in a halftone dot document. By counting the number of signals, the halftone dot area is determined.

FIG. 23-1 shows a determination pixel group for halftone determination. As shown in Fig. 23-1, characteristic density changes are detected in pixel groups each consisting of four pixels as indicated by thick frames at 2251, 2252, 2253, and 2254 with the pixel of interest at 225° as the center. Detect halftone dots.

FIG. 18-2 shows the configuration of the halftone feature extraction section 1827.

In FIG. 18-2, halftone dots are detected as follows using the density direction signal DSL from the character edge determination section.

The output of the gate 1851 is small (
The output of gate 1852 indicates that there is an upward density change of at least one pixel in pixel group 2253; The output of gate 1854 indicates that there is an upward density change of at least one pixel, and the output of gate 1854 indicates that there is a downward density change of at least one pixel in pixel group 2254. The output of gate 1856 indicates that there is a density change to the right for at least one pixel, the output of gate 1856 indicates that there is a density change for at least one pixel to the left in pixel group 2251, and the output of gate 1857 indicates that there is a density change for at least one pixel or more in pixel group 2252. This indicates that there is a leftward density change, and the output of gate 1858 indicates that there is a rightward density change of at least one pixel in the pixel group 225I.

These outputs are logically operated by gates 1859 to 1865, and the result is 2210°22 in Fig. 23-2.
The output DOTO becomes "1" in four cases as shown in 11, 2212, and 2213.

In Figure 23-2, the symbol → indicates that there is a rightward density change in one or more pixels in the pixel group surrounded by the thick line, and similarly, the symbol ← indicates that there is a density change in the right direction for one or more pixels in the pixel group surrounded by the thick line. The symbol symbol indicates that there is one or more pixels with a density change toward the left within the group of pixels surrounded by a thick line, and the symbol symbol indicates that there is one or more pixels with an upward density change within the pixel group surrounded by a thick line. indicates that one or more pixels have a downward density change in the pixel group surrounded by the thick line.

In the case shown in 2210, it is a halftone part as shown in 2214 or 2215, and in the case shown in 2211,
It is a halftone part as shown in 2216.2217, and 221
In the case shown in 2, it is a halftone part as shown in 2218.2219, and in the case shown in 2213, it is a halftone part as shown in 2220.
This is a halftone part as shown in 2221.

Figure 18-3 shows a signal that determines whether there is a point near the pixel of interest where DOTO = "1" by applying a judgment in a wide area to the DOTO signal generated in step 1827 shown in Figure 18-2. Form DOTI. This is a halftone dot area detection section.

1831 is a determination unit that determines whether or not there is one or more points with DOTO=“l” in the 4×3 window that includes the pixel of interest; if there is, it indicates “I”; otherwise, In this case, “0” is output as “DOTO”.1
8311 and 18312 are line memories, each of which provides a delay of one line, and these three lines of DOTO are input to the flip-flop 18313 at the same time, and the OR gate 1
8314, flip-flops 18315, 18316, and 18317 are each delayed by one clock, and their outputs are input to OR gate 18318 and DOTO
get '. At this time, for example, as shown in Figure 18-4, when 1 (Figure) and O (B) are mixed and output as DOTO in three consecutive lines, 1852 A logical OR is performed within a 3×4 window indicated by , and DOTO' is calculated.

With this operation, D
The OTO signal is converted to a relatively continuous DOTO' signal.

On the other hand, 1832 in FIG. 18-3 calculates the DOTO' signal over a wide area and generates a DOTO signal indicating whether or not the pixel of interest is in the halftone area.

18321 and 18322 are line memories, each of which is delayed by l life. 18323.18324
is a calculator.

As shown in FIG. 18-5, for the pixel of interest 1861 (pixel in the i-th sub-scanning line, j-th pixel in the main scanning), DOTO' is sampled every four pixels in the main scanning and every l lines in the sub-scanning. Before the l line (i-1st line),
In this embodiment, the following calculations are performed with N being an appropriate integer.

N=16) The above SUMLI, SUMRI, SUML2. SUMR
Outputs 2.

18325 and 18326 are adders that calculate the sampling sum SUML of DOTO' on the left side of the pixel of interest as SUML1+SUML2===OSUML, and calculate the sum of the sampling sums of DOTO' on the right side of the pixel of interest as
The sampling sum SUMR of SUMR1+SUMR2−
>Calculated and output as SUMR.

18327 and 18328 are comparators, 18329
1830 is an OR gate, and 1830 is a ROM table, which outputs a halftone determination slice level value T4 in response to a 4-bit SEG signal input as an address.

In this embodiment, since N=16, halftone dots are detected in two areas of 4N=64 pixels before and after the pixel of interest in the main scanning direction and 5 lines in the sub-scanning direction.

DOT is output as "1" only when at least one of SUML>T 4 or SUMR>T 4 is satisfied, and becomes "O" otherwise. As a result, the signal DOT becomes a region signal that becomes "1" in the halftone dot region.

The contents of the ROM table 1830 are shown in FIG. 18-6.

As the SEG signal value increases, the slice level for halftone dot determination becomes smaller (becomes smaller), and even if a density pattern such as that shown in FIG. 23-2 is slightly present in the document, halftone dot determination is performed and a DOT signal is output.

That is, in the above embodiment, in order to determine a halftone image, it is determined whether or not there are points in which dots are connected like halftone dots in a 4×3 window, for example, and a predetermined number of points in which dots are connected in a predetermined area is determined. If there is more than that, it is determined to be a halftone dot area.

What has been described above is the feature extraction unit shown in the first diagram 403. The next pixel-by-pixel color determination signal SL from this feature extraction unit,
UNK, COL, CAN and character edge judgment signal EDGE
A color signal processing unit 402 and a color processing control signal generation unit 4 using
The operation of 04 will be explained using the color processing circuit shown in FIG.

Reference numeral 103 denotes a light quantity signal-to-density signal converter, which converts R, G, and B signals in the O to 255 range to C, M, and Y signals in the O to 255 range using the following equations.

The black components included in these C, M, and Y signals are
04 is determined as follows.

K=min (C, M, Y) Four color density signals C, M, Y, K with this K added
is calculated by the following equation in order to remove the undercolor in the UCR/Mask unit 105 and to remove color turbidity of the developing material of the printer 202.

...Formula... Here a 11 'va 141 a 21 'a 2
41 a 31- a ur a 41 to a44 are predetermined masking coefficients for removing color turbidity, and ul+u2. u3 has components M, C, Y
is a UCR coefficient to be removed from the color component of . Here, M', C', Y', and K' are the control unit 40
One is selected by the 2-bit developing color signal PHASE from 1 and output as a 1 signal.

0.1 of PHASE signal. M', C corresponding to 2.3
'Y' and 'Ni' are selected.

112 and 113 are line delay memories, which delay the generation of the character edge determination signal from the feature extractor by 3 lines and 4 clocks, so that the vI times signal M signal is similarly delayed by 3 lines and 4 clocks.

On the other hand, the color determination section 106 is delayed by two lines and two clocks until it generates a determination output such as BL or UNK. In order to match the delay amount with the delay amount of the character edge determination unit 107, the signals BLI and UNKI are delayed by one line and two clocks by the line delay memory 120.

Generate C0LI and CANI.

[Weighting is done by the addition section]

Next, the operation of the weighting adder section 114 to 116 in FIG. 1 will be explained. Figures 24-1 to 24-7 show color determination signals and single character edge determination signals for rAJ characters read in various color portions. The determination signals of the cross section indicated by letter a in FIG. 24-7 are shown in FIGS. 24-1 to 24-6.

FIG. 24-1 is a diagram showing a timing chart of each signal when a black rAJ character is read as black, and shows an achromatic density signal (hereinafter referred to as ND times signal). The M2 signal delayed by 113 is delayed due to the blurring of the reading optical system.
−It is read with an accent compared to Figure 7. Further, the edge signal is formed in a shape that bulges out from the end of the character due to the succession of density changes of AK3 and AK7 described above. BL as a color judgment signal
Only the I signal is generated.

Here, the M2 signal and EDGE signal indicating the ND times signal are as follows.
Since a green color separation signal is used, the outputs from FIG. 24-2 onwards also show roughly the same output as in FIG. 24-1 except for the green characters. For green characters, M2 signal and E
No DGE signal is generated.

Figure 24-2 shows the case where an rAJ character composed of color characters is read, and the COLI signal indicating that it is a color and the presence of color pixels with a density higher than the white pixel around the white pixel are shown. The CANI signal shown is generated as shown in the figure.

FIG. 24-3 shows a case where an rAJ character composed of intermediate chroma characters is read, and a UNKI signal indicating intermediate chroma is generated.

Figure 24-4 shows the case where rAJ characters composed of black characters are read with color shift, and while the BLI signal is thinner than in Figure 24-1, the intermediate chroma signal UNKI due to color shift is present around it. occurs. Figure 24-5 shows the case where rAJ characters composed of colored characters are read with color shift.
Compared to Figure 2, the C0LI signal is thinner, but there is a UN on the edge of the character.
A KI signal is generated. Furthermore, as the CAN1 signal is determined to be colored, the portion corresponding to the outside of the character edge becomes thinner.

FIG. 24-6 shows a case where a color character with near intermediate saturation is read with a color shift and a black determination pixel occurs at the edge.

In this case, the same signals as in FIG. 24-5 are generated except that the BLI signal is generated instead of the UNK1 signal.

Also, Figures 25-1 to 25-3 are the black letters in Figure 24,
a for each case of medium chroma characters, middle chroma characters on the edges of black characters
This is an enlarged cross-section. Here, v2 has a development color of M,
An example of the output signal of the circuit 105 in the case of C, Y, and Bk is shown.

Figure 25-1 shows the case when the black text is read, and the circuit 105
Since UCR is acting on the M, C, Y color components are 2
It has decreased to around 0%. However, since this character is a black character, it is desirable to record it using black toner as much as possible.

Further, it is desirable that the intermediate saturation occurring at the edges of black characters as shown in FIG. 24-4 cover the M, C, and Y color components as much as possible. On the contrary, it is desirable to reduce the intermediate saturation component that occurs at the edges of colored characters as shown in FIG. 24-5. Further, it is desired to distinguish the black component generated at the edge of a color character as shown in Fig. 24-6 from the black character edge shown in Fig. 24-i.

From the above, in this embodiment, UCR/M is determined according to the results of the color determination signal and character edge determination signal as shown in FIG.
Color recording signal amount 2 from the a sk circuit 105 (
M'C', Y', K') and the ND signal M2 are appropriately mixed to perform color recording.

In Fig. 26(a), it corresponds to the black character EDGE in Fig. 24-1, and when the developed color is M, C, Y, an O signal (no development) is output, and when the developed color is Bk, the density signal M2 is output. Output. FIG. 26(C) corresponds to the intermediate chroma J-tsuji in FIGS. 24-3 and 24-5. In this case, in order to emphasize the black component of the edge, M', C, which is generated from 105 as the color recording signal v2, is used for the developed colors M, C, and Y.
When the developed color is Bk, a signal obtained by adding 50% of the color recording signal v2 to the density signal M2 is output. Figure 26 (
f) corresponds to the non-edge portion in black letters in FIG. 24-1. Here, in order to improve the signal connection with the edge portion recorded in Bk monochrome, M' of the color recording signal v2,
The C' and Y' components are reduced to 374, and 1/4 of the density signal M2 is added to 3/4 of the Y' component during Bk recording. Figure 26(b).

In (d) and (g), the above black emphasis operation is not performed by the CANI signal.

Next, changes in the image signal due to the calculation in FIG. 26 will be explained using FIGS. 25-1 to 25-3. Note that here V2 (M
) means the v2 output when PHASF,=O (7 zenta developed colors). V2 (C), V2 (Y), V2 (B
k) also has two outputs for cyan, yellow, and black, respectively.

In Figure 25-1, it is the black text part, and the part b' is the 26th part.
This is an edge portion corresponding to (a) in the figure. Here, M
, C, and Y recording signal amounts are O, and a density signal M2 is output as a Bk signal. Part C is (f) in Figure 26.
It is a non-edge part of the black part corresponding to , and the developed color M
, C, Y ■4 signals ■4 (M), V4 (C), V
4 (Y) is V2 (M), V2 (C).

V2 (Y) 17) 3/4, which is the value obtained by adding 3/4 of V2 (Bk) and 1/4 of M2 as the Bk (7) signal.

FIG. 25-2 is a medium chroma character, and portion d is an edge portion corresponding to FIG. 26(C). Here, V4 (M
), V4 (C), V4 (Y) are V2 (M), V2
(C), V2 (Y) (7) 1/2 and V4 (Bk
) is the sum of 1/2 of v2 (Bk) and 1/2 of M2.

Figure 25-3 shows a case where intermediate saturation occurs at the edge part of a black character. Edge part e is processed in the same way as part d, and non-edge part is subjected to the same processing as part C due to black determination (BL=1). be done. This reduces the color signal at the edges of black characters.

Multipliers 114, 115 and adder 116 are used in FIG. 1 to generate V44 times the value in FIG. 26.

Then, in the multiplication coefficient generation section 108, the BLI.

The multiplication coefficient GA of the multiplier receives the UNKI, C0LI, and CANI color determination signals and the character edge determination signal EDGE.
Generates INI and GAIN2.

As shown in FIG. 27, the multiplication coefficient generating section 108 is composed of a ROM, and as shown in the figure, BLI, UNKI, C0L
I.

CANI, EDGE 5-bit judgment signal and PHASE
It is input as an address and correspondingly outputs two gain signals GAINI and GAIN2 of 3 bits each.

FIG. 28 shows the relationship between the address and output of this ROM.

The gain signal here is the actual gain multiplied by 4, and multiplier 114. 115, the input V2. Multiplied by M3.

FIG. 29 shows details of the multipliers 114 and 115. The 8-bit image signal is processed by bit shift type multipliers 2901 and 2902.
are multiplied by 4 and 2, respectively. These are 3-bit gain signals GAIN (2), GAIN (1), GAIN
(0) is added by adders 2906 and 2907 selected by 2903, 2904 and 2905. After that, it is multiplied by 1/4 by a bit shift type divider 2908 and 2
A 55 limiter 2909 rounds all 9-bit data of 255 or more into 255 8-bit data and outputs the rounded data.

The color recording signal v2 and the density signal M2, weighted and added based on the color determination signal and the character edge determination signal as described above, are input to the spatial filter 117.

[Spatial filter section]

Figure 30 shows the spatial filter in this example (1 in Figure 1).
17) is shown. The spatial filter shown in FIG. 30 is an edge enhancement filter using a 3×3 pixel Laplacian filter, and the Laplacian multiplier can be switched between two types of 1/2.1.

3001 and 3002 are line delay memories, respectively.

Three lines of image signal V4. generated by this line delay memory. V42. V45 is delayed by l clocks by flip-flops 3003 to 3006, respectively.

The pixel of interest here is V43, and V41. V42. V44
゜V46 is a multiplier 3007 to configure a Laplacian.
3010 is multiplied by (-l) and each adder 3011.301
2°3013 is added. Further, the pixel of interest V43 is multiplied by 4 in a multiplier 3014 and added to the output of the adder 3013 in an adder 3015 to generate a Laplacian L. This Laplacian signal is multiplied by 172 in a multiplier 30I6. Adder 3
At 0I7, the pixel of interest V43 and L/2 are added to generate a weak edge emphasis signal El. The adder 3018 adds the pixel of interest V43 and the Laplacian L to generate a strong edge emphasis signal E2. These two types of edge-enhanced signals and the signal V43 of the pixel of interest itself are the control signal D.
FIL (1), DFIL (0) selected v5
Output as 5 times. DFIL (1) is 0 and DFI
When L (0) is 1, the weak edge emphasis signal E1 is selected, and when DFIL (1) is l and DFIL (0) is 1, the strong edge emphasis signal E2 is selected, and DFIL (0
) is O, the image signal V43 without edge enhancement is selected and outputted as v55 times.

This filter switching signal DFIL (1), DF
A filter control signal generating section generates a 2-bit DFIL signal consisting of IL(0).

In this embodiment, strong edge emphasis is applied to the edges of black characters so that the edges of black characters are output sharply. Further, edge enhancement is not applied to non-character edge portions in order to prevent the color tone from changing due to edge enhancement. The edge portions of characters with intermediate saturation and colors are configured so that while the edge portions are recorded sharply, weak edge emphasis is applied so that the change in color tone due to edge emphasis is not so conspicuous. Note that when the CANI signal is 1, the EDGE is not emphasized because it is a BLI signal or UNKI signal generated by color shift at the edge of a color character. FIG. 31 shows a circuit of the filter control signal generator 109 shown in FIG. 1 which controls the filter shown in FIG. 30. The logical formula is shown in FIG.

In the FILTER circuit 117, the pixel of interest is delayed by one clock from the l line, so the filter control circuit generator 109
The FIL signal from the 1-line memory 121 is delayed by 1 clock from the 1-line to become the DFIL signal. Similarly, the GAM signal from the gamma switching signal generating section 110 and the SCR signal from the screen switching signal generating section 111 are delayed by l line and l clock, and are converted into DGAM signal, DSCRSC.
Above.

[Gamma conversion section]

The gamma conversion unit 118 shown in FIG. 1 performs density conversion of the image. The gamma conversion section 118 is composed of a ROM as shown in Fig. 33, and the 8-bit (5) signal subjected to the frilter processing is input as an address of the ROM, and the corresponding gamma conversion output is output from the data terminal of the ROM as an 8-bit signal. It is output as a VIDEO signal. Furthermore, 2-bit DG is input to the address line along with the v5 signal.
As shown in FIG. 34, four types of gamma conversion characteristics can be selected by the AM signal.

In FIG. 34, when DGAM=0, it is a case where input=output, and is applied to a non-character edge portion. D
When GAM=1, as shown in the figure, for inputs from 0 to 255, 0 side and 255 side are both 0 for inputs corresponding to j-section.
and 255 outputs between them. This means that for nearby inputs that are low-density inputs, video with a lower density
o signal is output, and for high-density inputs near 255, a higher-density Video signal is output, emphasizing the density change of inputs near 128, which is an intermediate density, so the character edge can be recorded more sharply.

This DGAMI is applied to color text edges.

In the case of DGAM=2, the value of j in DGAM=1 is set to an even larger value k, and the character edges are recorded even sharper. However, since the linearity between input and output begins to collapse, color tone cannot be guaranteed.

Therefore, DCAM=2 is applied to mid-chroma character edges.

In the case of DGAM=3, the characteristic uses a value of l that is even larger than k, and is applied to black character edges where sharpness is required.

This gamma switching signal DGAM is generated by the gamma switching signal generating section 1.
10, the GAM signal is delayed by one line and one clock by a line delay 121. The gamma switching signal generating section 110 is composed of a ROM as shown in FIG. 35, inputs the color determination signal and the two-character edge determination signal as an address, and outputs the GAM signal as data. The contents of the ROM table are shown in FIG. As mentioned above, the black character edge part (EDGE=1°BL1=1) has CAM=3, and the intermediate saturation character edge part (EDGE=1.UNK=1) has CAM=2, but in both cases, there is a color shift. By B
CAN indicating that L1=1 or UNK=1
If there is an I signal, CAM=0 is set so that the character edges are not emphasized.

(PWM modulation section) The gamma-converted VIDEO signal is sent to the PWM modulation section 119.
It is converted into a pulse width signal at . Then, by controlling the lighting time of the laser 213 using the pulse width modulated signal, a copy output 406 with gradation density expression is obtained.

FIG. 37 shows details of the PWM modulation circuit used in the modulation section.

The VIDEO signal is converted into an analog image signal AV by the D/A converter 3701. Image signal CL synchronized with VIDEO signal
Screw clock CLK4 of K and twice its frequency is H3YN in toggle flip-flops 3702 and 3703.
The clock signal CLK4F and CLKF are synchronized with C and divided into 1/2, respectively, and converted into clocks CLK4F and CLKF with a duty of 50%. These two clocks are transformed into triangular waves by integrators 3704 and 3705, and then adjusted in wave height by amplifiers 3706 and 3707 to match the output dynamic range of the A/D converter.
compared to the signal. As a result, the AV signal is PW4 and PW
is converted into two pulse width modulated signals. After that, the selector 3710 selects PW4 and PW from the DSCRSC.
One of them is selected and becomes the laser drive signal LDR.

FIG. 38 shows the operation timing of this circuit. As shown in the figure, the triangular wave TRI4, which is obtained by integrating the clock 4F obtained by dividing CLK4 by 1/2, is a triangular wave with a period of one image pixel.This triangular wave changes approximately linearly over the entire output range of the D/A converter. Therefore, this triangular wave and analog image signal A
By comparing the AV signal with V, the AV signal is pulse width modulated with one image pixel section as one period, and becomes PW4. Similarly, TRI is a CL whose frequency is divided by 1/2 from the pixel clock CLK.
Since it is generated by KF, the AV signal is pulse width modulated by this TRI, with one cycle being a two-pixel section of the image, and becomes PW.

The PW4 signal pulse width modulated in one pixel period is recorded by the printer at the same resolution as the clock CLK. However, when an image is recorded using the PW4 signal, the basic density unit is as small as one pixel, so the gradation expression cannot be said to be sufficient due to the characteristics of the electrostatic photographic process used in the printer.

On the other hand, since the PW multiplied signal reproduces the density in units of two pixels, the gradation expression is sufficient, but the recording resolution is half that of PW4.

Therefore, in this embodiment, by controlling the DSCR according to the type of image, PW,! - Switch PW4 for each pixel. Specifically, PW4 is used for black character edge portions and intermediate saturation character edge portions that require high resolution.

PW is used for color character edge portions and non-edge portions to emphasize color tone. However, it has been experimentally confirmed that it is better to use PW4, which emphasizes the resolution of color character edges, even at the expense of color tone, for originals composed of thin color characters such as maps. The signal DSCR for switching between PW and PW4 is generated by the screen switching signal generator 11.
The SCR signal from 1 is delayed by 1 line and 1 clock by line delay 121. Details of the screen switching signal generating section 111 are shown in FIG. 39. In addition, the screen switching signal generation unit 1 is used when recording fine color characters with PW4.
11 is shown in detail in FIG. The circuits shown in FIGS. 39 and 40 are provided in parallel inside the screen switching signal generating section 111, and these two circuits are switched according to the mode input from the operating section 407. As a result, in the DSCR of FIG. 38, the portion corresponding to the black or intermediate saturation character edge portion (also the color character edge portion in FIG. 40) becomes LOW, and PW4 is output as an LDR signal only in this section. At this time, even if it is determined to be a character edge part, in the case of a character edge part with color shift (CAN1=1), the PW4 signal should not be used to prevent the quality of the recorded image from deteriorating due to the color shift being emphasized. It has become.

The laser drive signal LDR generated by the above processing is supplied to the printer 201 in FIG. 2. Then, in accordance with this signal, the semiconductor laser 213 is pulse-width modulated driven pixel by pixel, and the laser beam is caused to scan the photosensitive drum 217 in a line based on the result.

As a result, the recorded images output from the printer 201 are as shown in FIGS. 45-1 to 45-6.

The original drawings in Figures 45-1 to 45-6 are the same as the various characters in Figures 24-1 to 24-6.

Figure 45-1 is a black character image. As shown in the figure, the edge portion S and non-edge portion P of the recording surface are subjected to separate processing using the EDGE signal determined for the peripheral portion of the character and the BL7 signal determined for the entire character.

Since the edge portion S is a strong black edge portion, only black toner is recorded with the density signal M2, as shown in FIG. Furthermore, strong edge enhancement as shown in FIG. 32 is applied.

Further, the gamma conversion characteristic of DGAM=3 shown in FIG. 34 is used. As a result, the black character edge portion is recorded as a sharp image close to a binary image due to the strong edge emphasis of the monochromatic black toner and the gamma characteristic that has a sharp slope. Further, in the 8th part, the DSCR signal shown in FIG. 37 becomes Low, so the laser is driven by the PWM modulation signal PW4 of I pixel period, and therefore a high resolution recorded image for each pixel is obtained.

On the other hand, in the non-edge part P, Y, M, and C are printed by the v22 part that has been subjected to UCR/masking color processing.

Recording is performed using four BK developing materials, and since no edge enhancement is applied and a linear gamma conversion characteristic of DGAM=O is used, recording is performed with color tones and gradations that are faithful to the original. In addition, since the DSCR signal is High in the P section, the PWM modulation signal Pw with a period of 2 pixels is
The laser is driven by As a result, since the printer expresses the density every two pixels, a recorded image with high gradations can be obtained.

FIG. 46-1 shows a recorded image as a result of switching the laser drive signal between a pulse width modulation signal PW for every two pixels and a mid-pulse modulation signal PW4 for every pixel in the DSCR signal.

Shown in Figure 46-2. 46-2 is an enlarged version of part a' (corresponding to part a of the original picture) of the recorded image in Fig. 46-1.
It is a diagram.

Here, the diagonal line portion is a portion recorded with a pulse width modulation signal in units of 1 pixel as shown in 4601, and the solid portion is a portion recorded in a pulse width modulation signal in units of 2 pixels as shown in 4602. When the laser beam scans the cross-sectional view of a, an SCR signal corresponding to the edge of the character is generated, as shown in the figure.
The outer periphery of the character is recorded using a pulse-in-pulse modulation signal for each pixel. As a result, a sharp recorded image faithful to the original with less jagged edges around the characters is obtained.

FIG. 45-2 shows a recorded image for a color text original. As shown in FIG. 26(e), the edge portion U of the color character recorded image is masked with each developed color of M, C, Y, and BK.
Apply weak edge enhancement to the OCR color signal ■2,
Furthermore, the sharpness is slightly improved by the slightly higher gamma conversion characteristic of DGAM=1. Since the laser is driven by a pulse width modulation signal with a two-pixel period, the color tone (gradation) is faithfully reproduced, although the sharpness of character edges is inferior.

Further, the non-edge portion P is similar to that shown in FIG. 45-1. 45th
The edge portion t of the recorded images in FIG. 3 and FIG. 45-4 is an edge portion for intermediate saturation determination. This part is shown in Figure 26 (c
), the M, C, and Y developed colors use only half of the v2 signal, and the BK developed color uses a signal obtained by adding half each of the 2 signal and the M2 signal.

Weak edge enhancement is applied to this signal. Furthermore, since the gamma conversion characteristic of DGAM=2 is used, the signal change at the edge of a character is relatively rapid even if it is not at the edge of a black character. As the laser drive signal, a pulse width modulation signal of one pixel is used for the t portion only when the BK developing material is used.

Further, the color tone of the edge portion is maintained by recording the Y, M, and C developing materials with pulse modulation signals in units of two pixels.

As a result, the Y, M, and C developing materials reproduce the color tone of the t portion with intermediate saturation, and the BK developing material realizes sharp character edges.

In FIGS. 45-5 and 45-6, intermediate chroma components and black components are generated in the periphery of color characters due to color misalignment.

In this case, the 45th-1 to 45th
-4 All edge processing explained in Figure 4 is canceled. This prevents color shift components from being emphasized and recorded.

When the ATLAS signal is copied at 1, the EDGE signal shown in FIG. 45 is generated even with a minute density change. Therefore, even if the characters are thin or on a colored background, the black two-color characters and the medium saturation characters are sharply recorded as shown in Figures 45-1 to 45-4. Therefore, a sharp image can be obtained.

Next, the control unit 401 responds to the input from the operating unit shown in FIG.
The control operation for changing the ATLAS signal and the SEG signal will be described below.

[Map mode/Standard mode]

As mentioned above, the map mode is a mode that makes it easier to determine fine characters and characters with low density as character edges.

By selecting this mode, the control section sets the ATLAS signal to 1.

As a result, the character edge determination unit 107 is shown in FIG. 18-7.

As explained in FIG. 20-2, minute density changes are detected over the entire density range from white to black. To select this map mode, the operator presses the image creation key at 613 in FIG. (
Fig. 47 4702). Then, the control unit 4.01
As shown in the figure, the display on the liquid crystal display section is switched from the standard screen 4701 to the image creation mode setting screen 4703.

In 4701, 4713 displays the copy magnification;
4 displays the recording paper size, and 4715 displays the set number of copies.

Each time the operator presses the field key 608 in 4704, the cursor 4712 moves down one step.

When the key 608 is pressed three times, the display screen is switched to 4705, and when the key 608 is further pressed twice for a total of five times,
Move the cursor 4712 to the map mode setting location.

Here, the OFF display is displayed darkly on a bright background, and the ON display is displayed brightly on a dark background, indicating that the map mode is in the OFF state (not set).

Here, when the + key 611 is pressed as shown in 4707, the control unit 401 identifies that the map mode is OFF, sets the ATLAS signal to O, and returns the display to the standard screen as shown in 4711.

When the ward key 610 is pressed in 4705, the control unit 40
1 indicates that the map mode is in the ON state (set state), with the OFF display of the map mode being displayed brightly on a dark background, and the ON display being displayed darkly on a bright background.

Here, when the + key 611 is pressed as in 4709, the control unit 401 identifies map mode=ON, sets the ATLAS signal to 1, and returns the display to the standard screen of 4711.

When the ward key 609 is pressed in 4708, the control unit 40
1 returns the display to the state of 4705 and map mode is turned off
Indicates that the state has been returned. When the copy start key 602 is pressed on the screen 4711, the control unit 401 performs a copy operation using the ATLAS signal set as described above.

When the copy start key 602 is pressed in each of the display states 4705 and 4708, the control unit 401 changes the ATCAS signal according to the display state of the map mode described above, returns the display screen to 4711, and then starts the copy operation. do.

(Control of SEG signal) Next, in accordance with the keystrokes of the operating section shown in FIG. 6, the control section 401
The operation of controlling the SEG signal to the character edge determining section 107 will be explained with reference to FIG.

When the asterisk key 612 is pressed as shown in 4802 on the standard screen 4801, the control section 401 changes the display on the liquid crystal display section to the *mode setting screen 4803.

Here, when the operator presses the trouble key five times to select the text/photo separation level 6, the control unit 401 changes the display to 4805 and displays the cursor 4815 at the text/photo separation level position. Here, when the operator presses the trouble key 612 or the giant key 611, the display changes to a screen 4807 for setting the text/photo separation level.

The text/photo separation level is divided into nine levels as shown in the figure, and each display position corresponds to a respective SEG signal value. At the leftmost position, 5EG=O, and as the position of the cursor 4816 shifts to the right, the corresponding SEG signal value also increases by one, and when the cursor reaches the rightmost position, 5EG=8. In the display state of 4807, 5EG=4. 480
When the ward key 609 is pressed twice as shown in 8, the control unit 401
makes the display look like 4809 and recognizes 5EG=2.

In 4807, when the same key 610 is pressed three times as in 4810, the display changes to 4811 and 5EG=7 is recognized.

4812.4 in each display state of 4809 and 4811
When the J key is pressed as in 813, the control unit 401
Output the G signal value and return the display to 4814. 4807.
When the copy start key 602 is pressed in each of the display states 4809 and 4811, the control unit 401 outputs the SEG signal value, returns to the screen 4814, and starts the copy operation.

When ATLAS=0, when the SEG signal value is increased, the halftone determination slice level Tl of the halftone area determination section becomes smaller as shown in FIG.
However, when halftone dot determination is performed, edges are extracted as shown in FIG. 20-2, and the slice levels Tl, T2 and T3 become large, making it difficult to extract character edges.

As a result, the number of sharply recorded character edges in the recorded image is reduced, and processing suitable for recording soft overall photographs is performed. (Photo priority) Conversely, when the SEG signal value is decreased, Tl increases, and the halftone signal DOT
is less likely to occur, and Tl and T2°T3 become smaller, making it easier to extract character edges. As a result, the number of sharply recorded character edges in the recorded image increases, and even minute character information is sharply recorded. (Character priority) Even when ATLAS = 1 (map mode ON), S
TI, T2, etc. depending on the magnitude of the EG signal. T3 is ATLAS
It changes with the same tendency as when =0. Therefore, increasing the SEG gives priority to photos, and decreasing SEG gives priority to text.

When ATLAS=1, D as shown in Figure 18-1
The OT signal is ignored and further A is ignored as shown in Figure 20-2.
Compared to TLAS=O, each value of T1, T2, and T3 is about half or less, so even finer characters and No. 18-7
The characters in the colored background have also been extracted for the circuit shown.

In the above explanation, G is used as the luminance signal of the color image signal.
color separation signals are used. However, the character edge extraction means described in this specification is not limited to color reading signals, but can also be applied to reading signals of black and white document reading devices that do not perform color separation, such as facsimiles.

[Second Embodiment] The aforementioned character edge determination unit 107 detects character edges based on the level difference between the pixels before and after the pixel of interest.

However, the optical image formed on the CCD 210 is blurred due to a deviation in the setting position of the original image imaging lens 209 shown in FIG. Due to this blurring, there are cases where the same character document can be detected as a character edge and cases where it cannot be detected.

The COD of G in the best focus state of the imaging lens used in this example for an original that repeats black and white at pitches of about 0 m, 2 m, and m as shown in Fig. 49 (a).
The output has an MTF of about 85% (Figure (b)).

Looking at the variations during mass production, the average value of this MTF is approximately 55.
% (Figure (C)).

Furthermore, the worst value of MTF due to blurring is about 40% (
d).

In this example, the original is
Since it is resolved at 0 dots, 0.2 mm corresponds to about 3 pixels. That is, the pitch between white and black of the original in FIG. 49(a) is also about 3 pixels.

On the other hand, in character edge detection in this embodiment, as shown in FIG. 19, level differences are detected at a distance of approximately 2 pixels in order to check the level differences between pixels to the left, right, top, bottom, and diagonal of the pixel of interest. It turns out.

In this embodiment, it can be resolved to the extent possible using the sampling theorem. Using an original with a 3-pixel period that is slightly coarser than a 2-pixel period, the edge detection slice levels T, , T2 . T3 and the halftone dot area determination slice level T4 in FIG. 18-6 are variable.

The original in FIG. 49(a) is black with a density of 2.0,
In actual manuscripts, for example, some characters are recorded using black information with a density of 0.2, in which case the black level is approximately 170, and the range between black and white is 85 levels, which is 1/3 of the black information with a density of 2.0. becomes. In order to determine black information with a density of 0.2 as text, the control unit 401 sets the value of 1/3 of the amplitude value W obtained by measurement to the center value of the text/photo separation level in 4807 of FIG. Generate SEG signal.

In the case of MTF 55%, W = 145, so T, = 1/
In 3W and ATLAS=0, 5EG=6, which is closest to T,=48, is made to correspond to the center value of the text/photo separation level.

In the case of MTF 85%, W=215, so T.

=71. Therefore, T, = which is closest to T, = 71
5EG=8 corresponding to 80 is made to correspond to the center value of the text/photo separation level.

When MTF=40%, W=105 and T,=3
5, so 5EG=4 is made to correspond to the center of the text/photo separation level.

In this embodiment, the SEG value corresponding to the center value of the text/photo separation level is limited to five levels from 5EG=4 to 5EG=8, and each of them is set from CENTER=O to CENTER=4.
5 levels of CENTER values are made to correspond to each other.

In Fig. 52, when the CENTER value is from 0 to 4, the 48th
The SEG values selected by the control unit 401 are shown in correspondence with the display scale of the text/photo separation level in FIG. 4807. Here, text/photo separation level 1 indicates the left edge of the display, giving priority to text. Level 9 indicates the right end of the display scale and gives priority to photographs.

FIG. 53 shows the input flow of the CENTER value by the factory or service personnel.

In this example, 5301. 5302 and 5303 are measured and calculated by the operator, but may also be done automatically.

The operator uses the numeric keypad 604 and the asterisk key 61
Enter 2 to enter text/photo separation level center value input mode. Recognizing this mode input, the control unit 401
05 is displayed on the liquid crystal display section 601. The operator inputs the CENTER value calculated in 5306 (the figure shows the case where 2 is input).

The control unit recognizes the CENTER value manually entered using the numeric keypad and displays it as 5307.

This CENTE is operated by the operator's short key 611.
The R value is stored in a nonvolatile memory (not shown). In a normal copy operation, the control unit 401 selects the SEG value corresponding to the text/photo separation level input based on the CENTER value shown in FIG. T
3 generator 2023.

The configuration of the second embodiment is almost the same as that of the first embodiment, and the values of the table 1830 are as shown in FIG. 51, and the contents of the table 2023 are as shown in FIG. 50. Only things are different. In Figure 50, ATL
The TI-T3 value in the case of AS=1 is a value determined experimentally to be more than half of the value in the case of ATLAS=0.

The T4 values in FIG. 51 are T, , T2 . T3. The value is determined to be almost the same as the value of T4, and is a value determined experimentally as in the first embodiment.

In this way, in the second embodiment, since a means for varying the character edge detection level in response to blurring of the optical system is provided, the same character edge determination signal EDGE can be generated even between devices with different MTFs of the optical systems. It becomes like this.

Note that the present invention adds a memory for storing one line of the G signal to the configuration shown in FIG. This also includes adding means for causing the G signal from the original to be stored in the memory, and having the control unit 401 automatically set the CENTER value shown in FIG. 53 using the maximum and minimum values of the signal values.

As explained above, according to this embodiment, by providing a means for varying the determination level of character edges, discontinuities in density in photographic recorded images caused by misjudgment of photographic originals as text originals and sharp high It becomes possible to suppress the occurrence of density dots.

Furthermore, it becomes possible to record clearly even thin text information and fine text information.

Furthermore, it becomes possible to clearly record characters on colored backgrounds such as maps and characters on halftone dots.

Further, it is possible to compensate for non-uniformity in character edge determination due to variations in MTF of document reading imaging lenses between apparatuses.

[Third Example] The character edge determination unit 107 in the first example uses the EDG
1903 in FIG. 19 for areas not determined as E.
This also includes halftone originals such as those shown in . CO
When reading pixel by pixel with D, moiré patterns occur due to the regularity of the CCD pixels and the regularity of the halftone original. In order to prevent this, in this embodiment, the FILTER circuit 117 is configured to apply smoothing to document areas that are not determined as character edges (areas that are likely to be halftone dots). As a smoothing filter, the pixel of interest is multiplied by 1/2 as shown in Fig. 41, and the surrounding pixels are multiplied by 1/2.
A smoothing filter is used that multiplies the pixels by 1/8 and adds them.

FIG. 42 shows details of the FILTER circuit 117 in this embodiment.

Here, it is connected to the A input of the selector 3020, and the third
Instead of the pixel of interest V43 selected under the conditions shown in FIG. 2(c), the SMG signal passed through the smoothing filter shown in FIG. 41 is selected.

Adders 4201, 4202, and 4203 add four pixels around the pixel of interest, V41. V42. V44. V46 is added. A signal V43F obtained by multiplying the pixel of interest V43 by four is added to the signal by an adder 42o4.

The result is 1/1 by a bit shift type multiplier 4205.
8, a smooth filter signal SM can be obtained.

[Fourth Embodiment] This embodiment is the PWM modulation section shown in FIG. 1, in which the developed color using a pulse width modulation signal of one pixel period is limited to Bk (black).

As mentioned in the first embodiment, sharp character edges are required for black character edges, and in the case of colored character edges, reproduction of the color tone of the original is more important.

On the other hand, as shown in FIG. 25-1, there are no M, C, and Y toners in the black character edge portion. Further, due to the action of the UCR circuit 105, almost no Bk) color exists in color characters. In addition, in the middle saturation character edge part, Bk) colors are also used as M and C, as shown in Figure 25-2.

Y toner is also present in moderation.

In consideration of the above characteristics, in this embodiment, the character edge determination section is made to be able to use the pulse white modulation signal PW4 of one pixel period for laser driving only when Bk toner is used.

As a result, the edges of black characters, which originally have few color components, can achieve the same sharpness as in the first embodiment, and the edges of color characters, which contain a small amount of color components, can be recorded sharply with only the Bk component, and the color components are Since gradation is maintained, color reproducibility is also guaranteed.

FIG. 43 shows the color processing circuit used in this example.

This circuit corresponds to that shown in FIG. 1, and the PHASE signal is input to the screen switching signal generating section 4301. 44th
The figure shows a screen switching signal generation section 4301 in this embodiment.
Shown in detail.

The gate 4401 decodes that the 2-bit PHASE signal is 3, that is, the developed color is Bk.

Then, it is used as an output permission signal of the NAND gate 4402. The other gate parts in FIG. 44 are the same as in FIG. 40,
As a result, the SCR signal becomes O only in the case of Bk development color at the character edge portion.

As explained above, according to the present invention, by simultaneously performing character edge determination and saturation determination of a document, it is possible to improve the sharpness of achromatic character areas while maintaining the color tone of color character areas, and to improve the sharpness of achromatic character areas and It is possible to remove color turbidity, improve the sharpness of text while suppressing the occurrence of moiré in the halftone dot area, and increase the amount of black material in black text, making it possible to reproduce clear black text. effective.

[Fifth Embodiment] In the first embodiment, it was described that the intermediate saturation determination signal UNK and the black determination signal BL are generated due to color blurring around color characters caused by scanning speed unevenness and imaging magnification error. .

The present invention is intended to find black needles and intermediate chroma areas of a document, use more black toner when recording those areas, and record sharper black or intermediate chroma images. be.

Therefore, if a UNK signal or a BL multiplied signal is generated due to the above-mentioned erroneous determination due to color bleeding, a large amount of black toner is used at the edges of color characters in a recorded image, resulting in a sharp image.

In order to prevent this, in the first embodiment, a CAN signal is generated by detecting the presence of a color signal (COL) with a small light amount value around the pixel of interest.

Even if the pixel of interest has intermediate saturation or a black signal, it is determined that this is due to color blurring around the colored characters, as shown in Figures 24-5 and 24-6. 26
The processing shown in the table in the figure is carried out to prevent black toner from being used in large quantities.

In the first embodiment, a G signal is used to detect the light amount signal. However, the G signal obtained by reading a green original shows the same maximum light amount value as a white original. Therefore, the color shift component that occurs around the green character has a smaller signal value in the G signal than the green character component,
CAN signal is not generated. As a result, a large amount of black toner is used around the green characters in the recorded image, deteriorating the recorded image.

Therefore, in this embodiment, instead of the G signal for detecting the light amount signal,
It uses a light intensity signal that does not depend on color.

FIG. 55 shows a CAN signal generating section in this embodiment.

FIG. 55 corresponds to FIG. 17-1 in the first embodiment. Then, instead of the G signal in Fig. 17-1, an ND double signal is generated, and the 3-line memory 1718°17
19 and 1720, the G2 signal, G3 signal, and G4 signal are generated by delaying the ND times signal by l lines each. This G2. G3. Each G4 signal is input to the same calculation unit 1722 as in the first embodiment to generate a CAM signal.

Here, the ND multiplied signal is a signal indicating the brightness of the original that does not depend on the color tone, and after the color separation signals R, G, and B of the original are each reduced to 1/3 by multipliers 4501, 4502, and 4503, they are multiplied by an adder. They are generated by adding them together in step 4504. Since the ND multiplied signals R, G, and B are added together at a ratio of 1/3 each, it can be said that the signal has all color components.

By using this ND multiplied signal as a brightness signal, the CAN signal sent from the calculation unit 1722 is generated for color turbidity occurring around color characters of all hues.

As a result, as shown in the table in Figure 26, the intermediate saturation and black judgment caused by color turbidity are canceled,
Black toner is no longer used around color characters.

[Sixth Embodiment] FIG. 56 shows an example in which turbulence affects two pixels when reading color separation signals. In the figure, a black signal corresponding to one pixel is generated at the outer edge of a color character due to color misalignment during reading. Furthermore, intermediate saturation occurs at the outer edge due to a slight color shift component.

1st. In the embodiment shown in FIG. 2, it is possible to generate a CAN signal that cancels the intermediate saturation determination or black determination for up to one pixel surrounding the pixel where the color determination signal COL is generated.

However, the UNK signal two pixels outside the COL signal shown in FIG. 56 remains because no CAN signal is generated. As a result, since the CAN signal is also generated in the portion where the black determination signal BL is generated, the M, C, and C signals are generated as shown in FIG.
, Y development color and Bk development color are both U CR/ Ma
The color signal v2 generated by the s k circuit 105 is recorded. On the other hand, since the CAN signal is not generated in the part where the intermediate chroma signal UNK is generated, the M
, C, and Y, only half of the color signal 2 generated by the UCR/Mask circuit 105 is used.
The density signal M2 is added to the Bk developed color. As a result, a larger amount of Bk toner may be used in the intermediate saturation judgment area at the outer edge than in the black judgment area at the inner edge, and in this case, a recorded image may have a black border around two dots of color characters. will be formed.

Therefore, the sixth embodiment is configured to generate a CAN signal that cancels the BL multiplied signal UNK signal generated around the two dots of the color determination signal. Its configuration diagram is shown in FIG. 51.

This figure is a substitute for FIG. 17-3 in the first embodiment. As in the fifth embodiment, the NDI signal, which is the average value of the R, G, and B signals, is output from the adder 45o4. This light amount signal NDI is stored in line memories 4701 and 4.
702, 4703, and 4704, the light intensity signals for 5 lines are delayed by l lines NDI, ND2°ND3.
ND4. Get ND5. This light amount signal is sent to the calculation unit 4705.
is input. Also, at this time, 5 lines of color determination signals C0LI corresponding to each light amount signal are simultaneously generated.

COI, 2. C0L3. C0L4. C0L5 is the calculation unit 4
Entered at 7o5.

FIG. 58 shows details of the calculation section 47o5.

Light amount signals for 5 lines NDI, ND2. ND3. ND4
゜ND5 and color determination signals C0LI, C0L2. C0L3
゜C0L4. C0L5 is each flip-flop 4801
~4812 causes a maximum delay of 4 clocks. Here, the pixels of interest are ND33 and C0L33. First, the first
Comparator 4813.4814.48 as in Figure 7-2
15°4816 and AND gate 4817.4818.
4819 and 4820, it is determined whether there is a pixel around the pixel of interest that has a smaller amount of light (higher density) than the pixel of interest and has been color determined. This completes the check around one pixel of the pixel of interest.

Next, comparator 4812.4822.4823.4
824 and AND gate 4825. 4826.482
7.4828, it is determined whether there is a pixel further one pixel outside the pixel one pixel outside the pixel of interest that has a small light amount value and has been color determined. This determines that two pixels having the characteristic of color shift exist outside the pixel of interest.

Furthermore, the light intensity level of the pixel of interest and the pixels around it are compared, and if it is determined that the pixel of interest has a larger light intensity value (lower density), the pixel of interest is influenced by the color judgment pixels two pixels outside. This means that it is a pixel that may have been misjudged.

To see this, the outputs of AND gates 4825-4828 and the outputs of comparators 4813-4816 are AN
A match is made by D gates 4818-4832.

For example, the pixel ND23 is one pixel above the pixel of interest, and
The pixel ND13 above the pixel is compared in size by a comparator 4821. If ND13 has a lower light amount value (higher density) than ND23, and ND13 has a color judgment pixel (CO
If L13=1), ND23 becomes a color misaligned pixel of color pixel ND13, and AND gate 4825 outputs I.

In addition, if the light intensity value of ND23 is smaller than that of the pixel of interest ND33, the pixel of interest ND33 is a color misaligned pixel two pixels outside of the color pixel ND13 (AND game) 48
29 outputs l.

In this way, AND gates 4817 to 482 indicate that a high-density color pixel exists one pixel outside the pixel of interest.
An output of 0 and an AND game indicating that a color pixel with high density exists two pixels outside) Each output of 4829 to 4832 is ORed by an OR game) 4833 to 4837 and output as a CAM signal. .

This CAN signal is treated in the same way as the CAN signal in Figure 1,
As shown in FIG. 56, the cor signal is used to cancel the BL multiplied signal UNK signal two pixels outside the signal.

As described above, according to this embodiment, it is possible to distinguish between the color cloudy component included around the color edge portion of the document and the signal of achromatic color or intermediate saturation.

As a result, when recording images, more black toner can be used to sharply record the achromatic edges, but on the other hand,
This has the effect of making it possible to record images with high saturation without using unnecessary black toner in the color edge areas.

In addition, especially in this embodiment, R9G,
Since a signal that is a combination of the B signals is used, it is possible to effectively prevent, for example, black blurring around monochromatic green characters.

[Seventh Embodiment] In the embodiment shown in FIG. 42, smoothing processing was performed in all areas other than the character edge area.

Smoothing processing has the advantage of being able to reduce moiré of halftone dots, but also has the disadvantage of impairing the sharpness of the image.

The seventh embodiment improves this drawback, and as shown in FIG.
07, the halftone dot area signal DOTI is sent together with the character edge area signal EDGE, and the area is divided into four areas as shown in FIGS. 60 and 61 FiL (0).

Generate FiL (1). Note that FIGS. 60 and 61 are modifications of FIGS. 31 and 32, respectively. As shown in FIG. 62, the filter 117 has four types of characteristics, and by smoothing only the halftone dots, it prevents the image from losing its sharpness at areas other than the halftone dots.

[Eighth Embodiment] In the previous embodiment, only the edges of the characters were output in a single black color in consideration of the reproduction of black characters.

Even in the case of black halftone dots, it is possible to consider a method of faithfully reproducing the color tone (=gray balance) of the black halftone dots by outputting a single black color. Then the 60th multiplication factor
As shown in the figure. That is, for the embodiment shown in FIG.
By outputting in monochromatic black when DOT-“1” and BL1="l" as shown in (i), black characters and black halftone dot images can also be output in monochromatic black.

In the embodiments described above, the present invention has been explained using a color copying machine as an example, but the present invention is applicable not only to such a color copying machine but also to other devices such as a single scanner, and furthermore, The apparatus may have a single image processing section without a part of the scanner.

Further, in this embodiment, the image processing is switched by switching the spatial filter, changing γ, or switching the screen (number of lines), but the present invention may use these individual processes.

Further, in the present invention, an electrophotographic color printer is used, but the present invention is not limited to this, and the present invention can be applied to other printers such as a thermal printer, an inkjet printer, or a bubble jet printer.

[Ninth Embodiment] According to the ninth embodiment of the present invention described below, for example, Y (yellow), 2M (magenta), C (cyan), Bk
The sharpness of each developed color (black) is set independently, and appropriate sharpness values can be selected for each of M, C, Y, and Bk.

Furthermore, character sharpness and photographic sharpness are set independently, so that appropriate sharpness values can be selected for each of the text and photographic parts.

Furthermore, in view of the above problem, from an image signal having density,
A character edge area is detected based on the continuity of density changes in the pixel of interest and its neighboring pixels, and halftone dots are created separately from the character edge area by the presence of density changes in different directions around the pixel of interest. The process is selectively switched based on the detected result and the mode selected by the operator's will.

FIG. 67, which is an overall block diagram of this embodiment, is almost the same as FIG. 1 of the first embodiment, and descriptions of the contents of individual circuits that are also common will be omitted. As described later, M
The difference from FIG. 1 is that ODO and MODI signals are input from a control section 401 to a multiplication coefficient generation section 108, a filtered control signal generation section 109, and a screen switching signal generation section 111.

[Explanation of the operation section]

FIG. 64 shows details of the multiview of the operating section 1870 shown in FIG. 6.

Numeric keys 601 are used to input numerical values from 0 to 9 when inputting the number of copies, zoom magnification, etc.

Reference numeral 602 denotes a display panel using a liquid crystal display, and is used to inform the operator of the current machine setting mode, paper size, copy magnification, etc.

603 is a reset key, which is a key for initializing the currently set mode. For example, it is used when desired settings cannot be made due to an erroneous operation.

604 is a clear/stop key, which is used to stop the operation when the machine is in operation, and to clear the numerical value set with the numeric keypad, such as the number of sheets, when the machine is not in operation.

A copy start key 605 is used to start a copy operation.

606 selects the paper size, and the selected paper size (for example, A4) is displayed on the display panel 602.

A density key 607 is used to adjust the density of the copy from light to dark, and 608 displays the current density level using nine LEDs.

A document type mode selection key 609 is used to select text mode, photo mode, and text/photo mode depending on the type of document.

Reference numeral 610 indicates an LED, which indicates that each of the text mode, photo mode, and text/photo mode is selected, and only one of the three is lit.

Reference numeral 611 indicates control keys, which are composed of an OK key 612, an up arrow key 613, a down arrow key 614, a right arrow key 615, and a left arrow key 616. Used to move the cursor and set each mode.

Reference numeral 617 is an image create key, which is used when it is desired to process and output an image or when adjusting various image processing conditions.

Reference numeral 618 denotes an asterisk key, which is used to determine the magnification according to the paper size or to register the copy mode as described later.

619 is a color mode key, which selects four colors (Y.

Specify the color mode such as M, C, Bk printing), 3-color (Y, M, C printing), monochrome, etc.

Reference numeral 620 is a color selection key, which selects one of AC3 (black and white/color automatic recognition) mode, black (black and white printing) mode, and full color mode.

Reference numeral 621 denotes a color selection display section, in which a lamp of a color mode set by a color selection key is lit.

Reference numeral 622 is an area specification key, which is used when specifying an area using area specification means such as an editor.

[About sharpness]

In this embodiment, text portions and photographic portions of a document can be automatically determined or manually set, and image sharpness and smoothness can be specified independently for each. The functions are described below as "Character sharpness" and "Character sharpness".
65 shows the designation method and changes in display panels. Specifically, "sharpness" can be obtained by edge enhancement, and "smoothness" can be obtained by smoothing.

701 is a standard screen, and this state is displayed on the normal display panel 602.

When the image create key 617 is pressed here, various functions for image creation are displayed. Therefore, when the figure key 614 is pressed twice, the cursor on the screen descends one after another, and the cursor moves to the character sharpness line as shown in 703, where the character sharpness can be specified. That is, in this state, the stuttering key 615 and the garden key 61
By pressing 6, character sharpness can be specified in five levels from weak (1) to strong (5).

For example, in the state 703, the character sharpness is at the middle value (3), but by pressing the same key, the state changes to 704, where the character sharpness becomes (4), and in this state, if the En key is pressed, the state returns to 703.

When the same key is pressed in the state 703 or 704, the cursor moves to the photo sharpness line, and the photo sharpness can be specified using the same key and the garden key in the same way as character sharpness. (705, 706).

703, 704. 705. In each case 706, pressing the subject key 612 returns to the standard screen 701.

[Explanation of original mode]

FIG. 66 shows each document mode.

First, in the text mode, the main focus is on clearly copying a text document, and in the photo mode, emphasis is placed on the color and gradation of the document in order to realistically reproduce a photograph (including halftone dots).

In text/photo mode, in a document containing text and photos (including halftone dots), the text and photos are separated and the text is clearly reproduced and the photos are reproduced realistically.

Here, for most originals, if you copy in text/photo mode, the photo parts will be copied realistically and the text parts will be copied clearly, and there will be no problem, but for some originals, for example, , etc. In the case of crowded characters, etc., it becomes difficult to detect the character edges, and the character edges may be recognized as part of the photograph (halftone dots), making it impossible to reproduce them clearly.

On the other hand, ■If there are sharp edges in a photograph, they may be recognized as characters, and the edges may be unnaturally emphasized, resulting in an unsightly appearance.

Furthermore, if the background is a map with characters drawn in halftone dots, it may be detected as a halftone image and the characters may not be clearly copied.

Therefore, to deal with the disadvantages like ■, copy in character mode, all character sharpness is applied,
Problems like ■ are dealt with in photo mode, and all photo sharpness is applied. The problem described in (2) can be countered by using the map mode as described above. By selecting the original type mode according to the original, more preferable copies can be made.

FIG. 68 shows the edge area determination means 110 in this embodiment.
3 (FIG. 11) is shown.

In FIG. 68, 1801 is a density change point detection unit, 1802 is a density change continuity and halftone dot extraction means, 18021 is a part for determining the area of the extracted halftone signal, and 18022 is a part for determining the area of the extracted halftone dot signal. This is the part where the character edge area signal EDGE is finally formed by the continuous density change detected at 10821 and the halftone dot area detected at 10821.

Furthermore, 1871 is a CPU in the control unit in FIG.
Based on the operator's instruction through the operation unit 1870 (FIG. 1 407), the mode switching signal MODO as shown in FIG.
, MODI are generated, and MODO is sent to the multiplication coefficient generating section 108, filter control signal generating section 109, and screen switching signal generating section 111 according to the type mode of each original.

Send MODI.

Further, the CPU 1871 sends threshold data TI+T2+T3, which is optimally determined according to each mode, to the concentration change point detection section 1801.

FIG. 69 shows the configuration of the concentration change point detection section 1801. The structure and operation of FIG. 69 are roughly the same as those of FIG. 20-1. In this example, in Fig. 20-1,
Data T using ATLUS signal and SEG signal as address
,,T2. Instead of the ROM table outputting T3, in FIG. 60, T1.T3 is output directly from the CPU 1871. T2. This embodiment differs from the first embodiment in that T3 data is set in registers 2023, 2024, and 2025, respectively.

Here, the CPU 1801 sets the optimal combination of threshold data (Tr, T2.T3) for each of text mode, photo mode, and text/photo mode in registers 2023.2024.2025 according to the document type mode.

In this way, appropriate AKI to AK in each mode
Eight signals are output from the concentration change point detection section 1801. That is, by changing the value of TI+T2+T3, the strength of character edge detection, that is, the degree to which characters are captured, can be changed.

The above AKI to AK8 are density change continuous detection, halftone dot detection section 1
An EDGEO signal and a DOTO signal are generated through 802. Each signal is input to an edge determination section 18022 and a halftone signal area processing section 18021. The DOTO signal is converted to a DOTI signal at 18021. This point is similar to the case of the first embodiment described above.

FIG. 70 shows the EDGE signal forming section 18022 (FIG. 68).

1841.1842 is a line memory, and a total of 2
A line delay is applied to EDGEO, and the cycle is adjusted for two sub-scanning lines with DOTI. 1843 is composed of a flip-flop, and the main scanning period of EDGEO and DOT 1 is adjusted, and EDGEO' and DO
TI' is output.

1844 is an inverter, 1845 is an OR gate, 184
6 is an AND gate. Here, the ATLAS signal is CP
This is sent from U1871 and indicates that the mode is map mode.

When the ATLUS signal is present, that is, when not in the map mode, the DOTI signal is inverted by the inverter 1844 and becomes 1.
It is AND gated with the EDGEO' signal at 846. That is, if it is an edge (EDGEO' = 1) and there is no halftone dot (DOTI' = O), it is determined to be an edge (EDGE = 1).

On the other hand, when ATLAS=1, that is, in the map mode, 1845 is always High, and EDGEO'L/1846 is not written (AND game), and the influence of the halftone dot signal DOTI' disappears.

In other words, in the case of map mode (ATLAS = 1), the effect of halftone dot detection is removed in literature department determination, so even if there is fine text on a colored background like a map, the text will not be mistakenly judged as halftone dots. Clear characters can be reproduced.

Since such a map mode performs character detection processing by removing the effect of halftone detection, it is effective in a character mode for character determination and a character/photo mode.

This concludes the explanation of the determination unit 107 at the character edge.

[Multiplication coefficient generation section]

As shown in FIG. 71, the multiplication coefficient generating section 108
701 and AND gate 2702, number gate 270
3. Consists of an inverter 2704.

First, gate circuits 2704, 2703, 27021. : EDGE' = EDGE (MODI MODO
) is used to generate the EDGE' signal. That is, the EDGE' signal is (character mode or character/photo mode) and becomes 1 when EDGE=1. In other words, when in photo mode, always EDGE'
=O, otherwise E at the edge of the character
DGE'=1.

ROM27011:m is BLI, UNKI, C0LI
, CANI, and EDGE', and a PHASE signal indicating which toner is currently being developed with M, CY, or Bk, are input to the output address, and two corresponding 3-pixel toners are input. Gain signal GAINI
.

Outputs the GAIN2 signal as data.

[Spatial filter]

Next, a spatial filter used in this embodiment will be explained.

The spatial filter 117 clearly expresses characters by adding to the input image signal a product obtained by multiplying the weighting coefficients of surrounding pixels around the pixel (i.e., a convolution operation with a coefficient matrix). It is possible to perform edge enhancement and removal of high frequency noise, as well as smoothing such as removal of moiré in halftone images.

In the spatial filter in this example, FIG.
As shown in a), convolution is performed for the 7 pixels circled in a 5 x 5 window centered on the target Xi,i (1: number of pixels in the main scanning direction, j: number of lines in the sub-scanning direction). calculation is possible.

The coefficient matrix is shown in FIG. 72(b), and RO-
Four types of coefficients up to R3 can be set independently.

That is, (output) = ROX (X1j-2+ Xi, ++
2) + RI X Xi, 10R3x (X+-1
,1+X+++,H) + R2X(X knee 2,.

+　X　++2. i) Outputs the following output.

FIG. 72(c) shows a block diagram of the filter 117.

An input image signal V4 and a control signal DFIL to be described later are input, and a processed result V5 is output.

301, 302. 30:3. 304 is a FIFo memory, each providing a delay of l lines.

305, 306. ..., 317. Reference numerals 318 each designate a flip-flop, which latches input data at the rising edge of CLK, gives a delay of one pixel, and extracts the pixels necessary for the operation indicated by the circle in FIG. 72(a).

319, 320. 321. 322. 323.
324 is an adder, 325. 326. 327.
A multiplier 328 performs a spatial filtering operation.

329 is a coefficient generator, which generates spatial filter coefficients RO,
It generates R1, R2, and R3, performs the calculation shown in equation (1) as a result, and outputs the result to (5). In this example,
Since the coefficients can be finely adjusted up to RO-R3, fine image reproduction can be performed by spatial filtering. FIG. 72(d) shows a block diagram of the coefficient generator 329.

351 to 366 are 16 registers, and the control unit 40
1, ROO and RO are required in advance, respectively.
I.

. . , R33 is written and held by the CPU 1871. 367, 368. 369. 370 are selectors of 4 tol each. Figure 72(e)
It is composed of a gate circuit shown in FIG.
)) is 0. 1. 2.3, selectively outputs A, B, C, and D out of the four inputs.

Therefore, in the spatial filter section, the coefficient matrix can be switched in synchronization with the DFIL signal input in synchronization with the pixel of interest (=the pixel in question).

In this embodiment, for the filter switching signal DFIL, a register is set in advance so that when DFIL=0.1, the filter coefficient corresponds to photograph sharpness, and when DFIL=2.3, the filter coefficient corresponds to character sharpness. 351-3
56 to a predetermined value. In this embodiment, in order to obtain a coefficient matrix as shown in FIG. 72(h), the operator inputs the values shown in FIG. 72(g),
As a result, a set of coefficients corresponding to the sharpness value set by the CPU is set in each register. The operator can input a desired sharpness value from the operation unit 1870, and different values can be set for a plurality of areas by specifying the area as described later.

FIG. 72(h) shows selectable combinations of spatial filters. Horizontal direction is D FI L O ~ 4
, the vertical direction represents the strength of sharpness (which can be set in five levels in this embodiment). For example, if you look at sharpness 3, smoothing is applied with DFIL=O, and DFIL
When the value is 1, the edge is emphasized, and as it goes to the right as DFI L 2 and 3, the degree of edge emphasis becomes stronger.

Also, if we look at the case of DFIL=O, the sharpness l
The smoothing is the strongest, and as the sharpness reaches 23, the degree of smoothing becomes weaker. Sharpness 4 emphasizes edges, and sharpness 5 emphasizes strong sharpness and edges.

In this way, the larger the DFIL value and sharpness value, the stronger the degree of edge emphasis (lower degree of smoothing), and the smaller the DFIL value and sharpness value, the stronger the degree of smoothing (lower degree of edge emphasis). become weak.

Note that the above-mentioned filter control signal DFIL will be explained next.

[Filter control signal generator]

In the case of FIG. 73 of this embodiment, the configuration is basically the same as that of FIG. 31, but MODEO, MODEI and PHASE are
(0), PHASE (1), FIL
(0) FIL (1) The difference is that the signal is controlled.

That is, in the case of character mode, the mode signal is MODO
=0. Since MODI=1, FIL (0), FIL
The value of (1) is forcibly set to 1 regardless of the BL1 signal (black pixel) or the EDGE signal (edge), and in the case of photo mode, N0DO = l and MODI = O, so FIL (0
).

The value of FIL (1) is forced to O. In addition, in text/photo automatic discrimination mode, since MODO=MODI=1, the values of FIL (0) and FIL (1) are CANI, BLI, EDGE, UNKI, C0L as in Figure 31.
It will depend on I.

That is, in each document type mode, MODO = 0 in character mode, and FIL (0) = "1" at all times.
FIL (1) = "1", strong edge emphasis is applied to the entire copy, and character sharpness is applied.

When in photo mode, MODO="1", MOD1="O"
becomes N, FIL (0) = “ON or O’, F I
L(1)=“0”, and photographic sharpness is applied to the entire copy.

When in text/photo mode, MODO="1", MOD
I=“l”, FIL (0) and FIL (1) are switched according to the image area, and photograph sharpness and character sharpness are applied to each pixel.

In addition, in FIG. 73, PHASE (0), PHAS
E(1) is input and PHASE (0)=PH
When ASE (1) = 1, that is, in the case of a Bk toner phenomenon, and when FIL (0) = 0, the FIL (1) signal is always set to 1. This is because in the case of Bk printing, black pixels are reproduced more clearly and the image becomes more sharp when edge emphasis is applied more than in the case of other Y, M, and C printing.

Regarding C filter switching] First, the values of FIL (0), FIL (1) and FII
, is shown in FIG. 74(b).

Next, FIG. 74(a) shows the filter switching signal FIL and its application state in each document mode. In FIG. 74(a), in the case of character mode, FIL is used in the entire image area.
=3 is applied.

In the case of text/photo mode, FTL=3 is applied to the black text area (Bkl=1 and EDGE=1), and FIL=2 is applied to the intermediate color text area (UNK=1 and EDGE=1). , other flat parts (EDGE
= 0) and chromatic areas (COLI = 1 and EDGE = 1), when developing with Bk) toner (PHASE (0
) =PHASE (1) =l) has FIL=1,
When developing with C4M or Y toner, FTL=O is applied.

In the case of photo mode, FIL=1 is applied when developing with Bk) toner, and FIL=O is applied when developing with C, M or Y toner.

The value of FIL (DFIL) is 3.2. 1.
Processing is performed to emphasize sharpness in the order of 0, and to enhance the smoothing effect in the order of 0.1.2.3.

That is, the filter with FIL=3 is applied when the sharpness is most required, and is applied to the entire image area in the text mode and the black text portion of the document in the text/photo mode.

The filter with FIL=2 is applied when sharpness is required next to FIL=3, and is applied to the intermediate color text portion of the document in the text/photo mode.

FIL=O or FIL=1 is applied to the flat and chromatic areas in text/photo mode or the entire image area in photo mode, but the sharpness when developing with Bk toner is
, By increasing the sharpness higher than when developing with M or C toner, it is possible to add sharpness to the entire image, so when developing with Bk toner, FIL =
] is FIL when developing with Y, M or C toner.
=O applies.

As shown in the remarks column of FIG. 74(a), FIL=O.

When FIL=1, it corresponds to photographic sharpness, and when the sharpness value is switched to five levels using the operation section 1870, the coefficients of the filters for FIL=0 and FIL=1 are switched in conjunction with each other by the CPU 1871. Similarly, FIL=2. FIL=3 corresponds to character sharpness and is switched in conjunction.

In this way, there are two types of sharpness: photo sharpness and text sharpness.
The reason why the images are divided into two groups and linked together is that when FIL=0.1, the majority of images are determined to be photographic images, and when FIL=2.3, most of the images are determined to be text images. Since this is a case where a judgment has been made, it is better to change the sharpness strength in conjunction with this to prevent the reproduced image from becoming unnatural.

In addition, the setting of this sharpness strength is FIL20~
It goes without saying that each of the four parameters may be set independently. In particular, if you want to make text extremely sharp and reproduce photos extremely smoothly,
Such independent adjustment is effective.

[Screen switching signal generator]

FIG. 75 is a circuit showing the contents of the screen switching signal generating section 111. The basic configuration and operation are the same as those in FIG. 39, but in FIG. 75 MODI.

The difference is that the MOD2 signal is sent from the CPU 1871.

That is, the OR game shown in FIG. 75) 8006, AN
D gate 8001.8002.8003.8005, N
In the case of character mode by the circuit of OR gate 8004, MODI=l, MODO=O, and the output of 8001 is 0.8002 is 1.8003 is 0.80.
The output of 04 is always 0, and the output of 018005 is always 0, and 5CR=O. That is, it is always printed at PW4.

In addition, in the case of photo mode, MOD1=O, MODO=
l, and the output of 8001 is 0.8002 is 0.
The output of 8003 is 1. The output of 8004 is 1. The output of 8006 is always l, and 5CR=1. That is, pw is always printed.

Furthermore, in the case of text/photo mode, MODl=1゜M
ODO=1, the output of 5ooi is 0.8002, the output of 018003 is 1, the output of 0.8004 is 1, and S according to BLI, UNKl, CANI, EDGE.
A CR signal is generated in the same manner as in FIG.

In other words, in each document type mode, character mode (M
OD1=1. MODO=O), 5CR=“0”
”, always select PW4, for example, high resolution 400dpi (dot per 1nch)
is printed.

Further, in the case of photo mode (MOD1=O, MODO=1), 5CR="l" is always selected, PW is always selected, and printing is performed at, for example, a high gradation of 200 dpi.

On the other hand, text/photo mode (MODl=1.MODO=1
), PW4 is selected in the black character area,
In other cases, PW is selected.

In this way, an input signal enable circuit is formed when the laser driver for printing is driven separately for high resolution and high gradation according to the mode signals MODI and MODO,
Printing can be performed according to the type of image.

In this embodiment, the combinations of modes, determination conditions, and processing conditions are not limited to the above specific examples.

For example, in FIG. 75, in addition to the three modes in this embodiment, a map mode is defined separately, and MODO 20,
When MODI=O (other modes are the same as in the above embodiment), BLI in map mode.

Regardless of UNKI or CANI, in the case of EDGE-"I", 5CR=O and PW4 is selected, and the sharpness of characters peculiar to the map is maintained.

[Mode setting flow]

FIG. 76 shows the flow of mode setting in this embodiment.

When the mode selection key is pressed in step 4301 after the power is turned on, four document type modes (text, photo, text/photo, and map) are sequentially switched in step 4302.

When the copy start key is pressed in 4303, the MO mode is changed in 4304 as shown in FIG. 42-1 according to the document type mode finally selected in 4302.
The DO and MODI are set and the copy operation is performed at 4305.

[Processing flow]

FIG. 77 shows the processing flow in this embodiment.

After power on, if the original mode key is pressed in 4301, the original mode is switched in 4302, and if sharpness is set in 4306, the sharpness value is set and held in 4307. At this time, the sharpness settings can be set independently for text sharpness and photo sharpness.

If the start key is pressed at 4303, 43o2
43o4 corresponding to the original mode held in
MODO and MODI are set in , and RO is set in 43o8 according to the sharpness value held in 43o7.
The values of O to R33 are set in the register. Furthermore, 43
A copy operation is performed at 05.

As explained above, according to the above embodiment of the present invention, the image quality intended by the operator is achieved by varying the image processing according to the mode selected by the operator's will and the detected image judgment result. can do.

In particular, by changing the black character processing according to the mode signal, it is possible to finely handle photographs, text images, text/photo mixed images, etc., and perform optimal processing for each.

Note that the mode setting is not limited to the above example, and various modes such as a glossy mode and a matte mode can be considered. Also, the processing that changes depending on the mode is black character processing (
The present invention is not limited to character edge detection, halftone dot detection, etc.), and can be applied to various other image processing such as changing screen signals.

Furthermore, this embodiment has also been described using a digital color copying machine as an example, but as described above, it can be applied to various other image processing apparatuses.

Furthermore, according to this embodiment, the sharpness of M, C, Y and the sharpness of Bk can be selected independently.
- Appropriate image quality can be obtained by increasing the sharpness of Bk in particular.

That is, by making it possible to independently set the sharpness of a plurality of color component signals including Bk (black), it is possible to improve the image quality by, for example, creating an image with a black tint.

Furthermore, according to this embodiment, the sharpness of the text area and the sharpness of the photo area can be selected independently, so that it is possible to select the sharpness of the text area and the sharpness of the photo area independently. By making it possible to set the sharpness of the image continuously and independently depending on the type of image (e.g. / photographic mixed image, etc.), it is possible to obtain an image with appropriate sharpness depending on the purpose and preference.

Note that the sharpness strength can be set manually using the keys on the operation panel, or may be input from the keyboard when connected to a computer.

Furthermore, the range of settings is not limited to the specific example of this embodiment.

Furthermore, the size, format, and coefficients of the spatial filter may be changed freely.

(The following is a margin) Next, for images that are difficult to distinguish as text or halftone parts, areas are specified, and by controlling the discrimination means for the specified areas, the ability to distinguish text or halftone parts is achieved. An example that can improve this will be described.

[Operation Unit and Editor] FIG. 78 shows an elevational view of the apparatus of this embodiment. Reference numeral 702 denotes an editor that also serves as the document presser 200. Editor 7
By using the touch ben 704 to specify two diagonal points 705 and 706 of the area on the document 703 placed on the document 703, the area of the hatched open area can be specified.

Reference numeral 701 is an operation unit including keys and a display unit.

[About Area Designation] In this embodiment, up to four areas can be designated and the document mode can be selected independently for each area.

An example is shown in Fig. 79, 8o1 (indicated by ■), 80
The document mode can be independently selected for the four areas 2 (indicated by ■), 8o3 (indicated by ■), and 8o4 (indicated by ■) and the other areas shown by ■. Here is the circle number～
■ corresponds to the order set by the editor, and although there are overlaps in the areas of ■ and ■, and ■ and ■, priority is given to the area specified later, and as shown by diagonal lines in Figure 79, 4 of ■ to ■ One area is uniquely specified.

[About specifying the original mode] The original mode can be specified independently in the areas ① to ② shown in FIG. 79 and the other area ②, and the method for specifying this will be described below.

First, regarding area (2), by pressing the document mode key 609 in FIG. 6, text mode, text/photo mode, and photo mode are sequentially switched and displayed in 610. If the original mode key 609 is not pressed, the default areas (1) to (2) of the device are specified by specifying the area with the editor. This flow is 80th
As shown in the figure.

In FIG. 80, 901 is a standard screen, and when the area designation key 618 is pressed, a screen like 902 appears.

If the key is pressed here, 0 partial processing is selected as shown in 903, and the screen shown in 904 is displayed.

In the state of 904, the details of the partial processing are displayed, and the key is pressed 5.
Press twice to move the cursor to the original mode.
The screen will be 05. If you press the bow keys one after another, the document mode will switch cyclically as follows: text mode (905) -> text/photo mode (906) - photo mode (907) -> text mode (905) ->... Press the key to switch in reverse order. If you press the key in the state of 905, 906, 907, the screen of 910 will appear and you will be able to select the map mode. In this state, switching between 910 (map mode OFF) and 911 (map mode ON) is performed using the H key and the OFF key.

If the OK key 612 is pressed in the states 910 and 911, the screen 908 will appear, and in this state, in 909,
After specifying two points with touch ben 704, press OK key 6.
By pressing 12, one area is specified,
Return to the standard screen 901.

When specifying four areas from ■ to ■, it is sufficient to repeat the above operation four times.

Furthermore, 0N10F in the map mode corresponding to ■ in Figure 79
For F, after pressing the image create key 619 in Figure 6, press the key several times to display the screen 910 in Figure 80, then switch using the Minister key and the Cut key, and press the OK key to return to the standard screen 901. .

Furthermore, in text mode and text/photo mode, a new map mode can be set to make the text appear more clearly. (FIG. 44) [Mode Setting Flow] FIG. 81 shows the flow of document mode setting in this embodiment.

When an area as shown in FIG. 80 is set in 4306, the document mode within the area is set in 4037.

If the mode selection key is then pressed at 4301, 43
02 has four document type modes (text, photo).

text/photos, map) will change sequentially. When the copy start key is pressed in 4303, MODO and l are set as shown in Figure 42-1 in 4304 according to the document type mode finally selected in 4302, and the copy operation is started in 4305. is executed.

[Signal Flow] Next, the overall signal flow in this embodiment will be shown using FIG. 82. FIG. 82 is a modification of FIG. 4, and parts common to those in FIG. 4 are designated by the same numbers.

In FIG. 82, 406 is a CPU and an editor 70.
2 and the operation unit 701 in the manner described above to obtain the input of the area and the document mode in each area.

407 is an address generation circuit which generates a main scanning address (Y address) and a sub-scanning address (X address) according to a synchronization signal. 408 is an area signal generator which, in synchronization with document scanning, outputs MODO,
Generates MODI and ATRAS signals, and features extraction unit 403
MO outputs the ATRAS signal to the color processing signal control generation unit 404.
DO and MODI are input.

[Area Signal Generating Section FIG. 84 shows details of the address generating circuit 407 and the area signal generating section 408. 1001 and 1002 are up counters, which generate addresses corresponding to pixels in the main scanning Y and sub scanning X as shown in FIG.

1003 to 1006 are window comparators, and C
It is compared with a value set in advance by the PU to determine whether it is within the designated area.

FIG. 84 shows the inside of the window comparators 1003 to 1006 shown in FIG.
~10104 and registers 10105 to 10108, registers 10105 to 1 remaining from AND gate 10109
A value corresponding to an area designated in advance by the CPU is set in 010B.

That is, by setting the upper limit of main scanning in 10105, the lower limit of main scanning in YU 10106, the upper limit of sub-scanning in YL 10107, and the lower limit of sub-scanning XL in XU 10108, the output will be XL. It becomes 1 only when <X<XU and YL<Y<YU, and generates a signal indicating whether or not it is within a rectangular area.

In FIG. 83, 1003. 1004.1005.
1006 are respectively ■, ■, ■, ■ in Figure 79
Inverter 1008. 1009. 1010. Antgame) 1011.1012.1013 performs processing that gives priority to later specification, and as a result, A4. , A3. A2. A
l is "l" corresponding to each area of ■, ■, ■, ■ shown in FIG. 79, and there is no overlap.

Furthermore, 1014 is a register directly connected to the CPU, and the 79th register
ATRAS, MODO, and MODI corresponding to the modes in each of the five areas (■, ■, ■, ■1■) in the figure are preset, and inverters 1015 to 1018
．． 77 Dogate 1019-1023. OR game) 10
MODO corresponding to each area of ■, ■～■ by 24,
MODI and ATRAS signals are output.

[Other embodiments shown in Fig. 82] FIG. 86 shows another embodiment of the embodiment shown in FIG. 82.

In the embodiment shown in FIG. 82, four rectangular areas are specified, but the shape and number of areas are not limited to this.

In the embodiment shown in FIG. 86, which will be described below, regions of any shape can be specified without regard to the number of regions. 8th
In FIG. 6, 4601 is a CPU, 4602 is an editor, and together with an operation unit 4603, a non-rectangular area and mode can be set.

The result is written to bitmap memory 4605.

Note that the bitmap memory 4605 has 3 bits in the depth direction, and each bit is the above-mentioned MODO and MODI.

Compatible with ATRAS. 4606 is the same address generator as 407, and 4603 and 4604 are selectors. The portion shown in FIG. 46 corresponds to 408°406, 701, 702 in FIG. 82.

Initially, the CPU 4601 selectors 4603 and 460
By selecting 4 to the B side, bitmap memory 4
605 can be freely accessed, corresponding to a non-rectangular area specified by the editor 4062, and its modes MODO and MODI are stored in the bitmap memory 4605.

Expand and write ATRAS.

Next, the CPU selects the selectors 4603 and 4604 to the A side, and accesses the bitmap memory 46o5 with the address generated by the address generator 46o60 to read the MODO previously written.

MODI and ATRAS can be read.

[Other Embodiments] In the embodiments shown in FIGS. 82 and 86, the manuscript mode (1 character/photo, photo mode) and map mode are specified for each area, but the invention is not limited to these modes. , other parameters may be set.

The embodiments shown in FIGS. 87(a) and 87(b) are shown.

In FIG. 87, the same screens as in FIG. 80 are indicated by the same numbers. Similar to Figure 87, 0N10F in map mode
When the M key is pressed on the F selection screen 910 and 911, the screen moves to 4703 in FIG. ) The character sharpness specification decreases by one with the off key (4704 → 4703)
You can specify text sharpness in 5 levels.

When the key is pressed in the state 4703 or 4704, the cursor moves to photo sharpness, and similarly to 4703 and 4704, the photo sharpness can be specified in 5 levels using the down key and off key. (4705,4706)4705.40706
When the key is pressed in this state, the text/photo separation level can be set, and can be specified in nine levels using the H key and the R key. (4707°4708) If the OK key is pressed in any of the cases 4703 to 4708, the process moves to 908 in FIG. 87(a) and waits for area designation. (By repeating the above operations, an unlimited number of areas can be specified.) In the area marked ■ in FIG. 79, the text/photo separation level and sharpness can be specified by normal operations as outside the specified area.

Further, as shown in FIG. 88, the CPU 406 receives addresses from the address generator 407, and in synchronization with each address,
MODO, MODI, and ATRAS signals may also be generated.

Furthermore, the CPU uses ROO to
The value of R33 is sent to 402, and Tl, T2 . Send the value of T3 to 403.

The processing flow is shown in FIG. 89. In 4901, when an area is specified, in 49o2, the mode (one character/photo) is selected within the specified area.

(photo, map, sharpness, text/photo separation level) are specified.

When the mode is switched by normal key operation in 4903, the mode (
9 sentences/copy, map, sharpness, separation level) can be switched.

49o if the start key is pressed in 4905
7, during the copy operation, TI, T2゜T is synchronized with the specified area according to the text/photo separation level.
3 outputs MODO, MODI, and ATRAS according to the sharpness, and ROO-R 33 outputs MODO, MODI, and ATRAS according to the original mode.

In the embodiments described above from FIG. 78 onwards, by changing the criteria for text/photo discrimination within the specified area, even if there are areas in the target image that may result in erroneous detection for text/photo discrimination. Even if there is such a range, it is possible to perform good image discrimination by specifying such a range in advance and inputting a discrimination criterion that allows good discrimination to be performed within that range.

In addition, in this embodiment, the standard for character/photo discrimination is changed, but this is not limited to this, and text/photo discrimination can be used for specific areas.
The above-mentioned text mode or photo mode may be uniquely set without determining whether it is a photo. In this embodiment, the target image is a color image, and halftone color images are often processed. Therefore, by combining the above-mentioned area specification and text/photo discrimination control, extremely accurate image area processing is possible. It is possible to make a judgment.

In addition, in this embodiment, the character sharpness and photo sharpness set in the map mode can be set independently within the specified area as shown in FIGS. 87(a) and 87(b). For example, specify the area of text in an image where text and photos overlap and mix, such as in a product catalog, and set the text sharpness to be strong for this area, or use the map By specifying the mode, it is possible to obtain a desired image in which the edges of the text portion are emphasized.

〔Effect of the invention〕

As explained above, according to the present invention, since the discrimination conditions for the discrimination means for discriminating character parts or halftone parts in a specified area in a target image are set, erroneous discrimination may occur in the above-mentioned discrimination. For areas that are likely to be affected, good discrimination can be achieved by controlling the discrimination means, for example, by setting discrimination conditions that are suitable for such areas in advance.

(Margin below)

[Brief explanation of drawings]

FIG. 1 is a diagram showing the configuration of a circuit block according to an embodiment of the present invention. FIG. 2 is a diagram showing the configuration of a copying machine according to an embodiment of the present invention. FIG. FIG. 4 is a diagram showing the circuit block of the embodiment shown in FIG. 2. FIG. 5 is a diagram showing the waveforms of the clocks CLK and CLK4 shown in FIG. 4. FIG. 6 is a copy of the example shown in FIG. FIG. 7 is a block diagram showing the configuration of the area processing section shown in FIG. 11 shown later. FIG. 8 is a diagram explaining the operation of the blocks shown in FIG. 7.
10 is a diagram showing the state of color bleeding, FIG. 11 is a diagram showing the configuration of the color determination section 106 and character edge determination section 107 shown in FIG. 1, and FIG. 12 is a diagram showing the R, G, B of the sensor 210. FIG. 13 is a block diagram showing the configuration of the pixel color determination unit 1101 in the color determination unit 106 shown in FIG. 11, and FIGS. 14-1 and 14-2 are Block diagram showing the configuration and operation of the iN detection circuit, Figure 15-1, Figure 1
5-2 is a diagram showing the configuration and operation of each selector shown in FIG. 13, FIGS. 16-1 and 16-2 are diagrams explaining the operation of the pixel color determination unit 1101 shown in FIG. 13, and FIG. 17- FIG. 1 is a block diagram showing the configuration of a CAM signal generator included in the area processing section shown in FIG.
2 is a block diagram showing the configuration of the calculation section 1722 shown in FIG. 17-1, FIG. 18-1 is a block diagram showing the configuration of the character edge determination section 107, and FIG. 18-2 is the same as FIG. 18-1. The halftone feature extraction unit 18 shown in FIG.
18-3 is a block diagram showing the configuration of the halftone dot area determination unit l shown in FIG.
828; Figures 18-4 and 18-5 are diagrams for explaining the operation of the circuit shown in Figure 18-3; Figure 18-6 is the table shown in Figure 18-3. 1830, Figure 18-7 is the signal conversion table 1 shown in Figure 18-1.
826; FIG. 19 is a diagram explaining the operation of the character edge determination section;
0-1 is a block diagram showing the internal configuration of 1805 shown in FIG. 18, FIG. 20-2 is a diagram showing the relationship between the input address and output data of the table 2023 shown in FIG. Figure 22-1 is a diagram showing typical dot arrays showing the patterns 1905 to 1912 shown in Figure 19.
Figure 22-2 is a diagram showing a detection pattern for detecting the dot arrangement shown in Figure 1, Figure 22-2 is a diagram showing a pattern at the end of a character, Figure 23-
1 is a diagram showing the state of halftone determination, FIG. 23-2 is a diagram explaining the operation of halftone determination, FIG. 24-1, FIG. 24-2,
Figure 24-3, Figure 24-4, Figure 24-5, Figure 24-
6 and 24-7 are diagrams showing the output of the feature extraction unit 403 when various characters are read, and FIG. 25-■, FIG. 25-2, and FIG.
4-1, FIG. 24-3, and a partially enlarged view of FIG. 24-4, FIG. 26 shows the multiplier 114 shown in FIG. 1. 115, a diagram showing the operation of the adder 116 and the multiplication coefficient generation section, FIG. 27 is a diagram showing the configuration of the multiplication coefficient generation section 108 shown in FIG. 1, and FIG. 28 shows the input address and output of the ROM shown in FIG. 27. FIG. 29 is a diagram showing the configuration of the multiplier shown in FIG. 1, FIG. 30 is a diagram showing the internal configuration of the filter 117 shown in FIG. 1, and FIG. 31 is a diagram showing the filter control signal shown in FIG. 1. 32 is a table showing the logical formula of the gate circuit shown in FIG. 31, FIG. 33 is a diagram showing the configuration of the gamma conversion section 118 shown in FIG. 1, and FIG. FIG. 35 is a block diagram showing the configuration of the gamma switching signal generating section 110 shown in FIG. 1. FIG. 36 is a diagram showing the relationship between the input and output of the ROM shown in FIG. FIG. 37 is a block diagram showing the configuration of the RWM modulation section 119 shown in FIG. 1, FIG. 38 is a timing chart for explaining the operation of each block shown in FIG. FIG. 40 is a block diagram showing the internal details of the screen switching signal generating section 111 shown in the figure. FIG. 40 is a block diagram showing the internal details of the screen switching signal generating section 111 when recording fine color characters. FIG. 41 is a pixel of interest. 42 is a diagram showing another configuration example of the filter circuit 117 shown in FIG. 1, and FIG. 43 is a diagram showing another example of the configuration of the filter circuit 117 shown in FIG. A diagram showing a configuration example, FIG. 44 is a screen switching signal generation section 430 shown in FIG. 43.
1, Figure 45-1, Figure 45-2, Figure 45-3, Figure 45-
4, FIG. 45-5, and FIG. 45-6 are diagrams corresponding to FIGS. 24-1 to 24-6, respectively, and include a timing chart showing the characteristics of each detection signal, and FIG. , No. 46
Figure-2 is a diagram showing further details of Figure 45-1, Figures 47 and 48 are diagrams showing display examples of the operation section shown in Figure 6, and Figure 49 is a diagram showing the C0D201 output 17) MTF. 50 is a diagram showing another example of the contents of table 2o23 in FIG. 20-1, FIG. 51 is a diagram showing another example of the contents of table 1830 shown in FIG. 18-3, and FIG. 48 is a diagram showing the SEG value selected by the control unit 401 corresponding to the display scale of the text/photo separation level 4807. FIG. 53 is a diagram showing the input flow of the CENTER value.
4 shows other embodiments of FIG. 1, FIGS. 55 and 57.
56 is a block diagram showing another embodiment of FIG. 17-1, FIG. 56 is a diagram explaining the operation of the embodiment of FIG. 57, and FIG.
FIG. 8 is a block diagram showing another embodiment of FIG. 17-2. Figures 59, 60, 61, 62, and 63 are views showing modifications of Figures 1, 31, 32, 42, and 26, respectively; Figure 64; 65 is a diagram explaining the sharpness setting, FIG. 66 is a diagram explaining the mode switching signal, FIG. 67 is an overall block diagram of the ninth embodiment of the present invention, and FIG. A basic block diagram of the ninth embodiment of the present invention, FIG. 69 is a circuit diagram of the density change point detection section, FIG. 70 is a circuit diagram of the edge determination section, FIG. 71 is a circuit diagram of the multiplication coefficient generation section, 72 is a diagram explaining the spatial filter, FIG. 73 is a circuit diagram of the filter control signal generation section, FIG. 74 is a diagram explaining filter switching, FIG. 75 is a circuit diagram of the screen switching signal generation section, and FIG. 976 is a circuit diagram of the filter control signal generation section. Mode setting flowchart, 7th
FIG. 7 is a flowchart for setting sharpness values, FIG. 78 is a diagram showing an operation section for specifying an area, FIG. 79 is a diagram showing an example of an area specified by the operation section shown in FIG. 78, and FIG. 78 is a flowchart showing the operation procedure of the operation section shown in FIG. 78; FIG. 81 is a flowchart showing the operation input procedure of the operation section shown in FIG. 80; FIG. 82 is another example of FIG. A block diagram showing the configuration of the device having the operation section, FIG.
2. Address generator 407 shown in FIG. 2, main address generator 40
84 is a block diagram showing the internal configuration of digitizers 1003 to 1006 shown in FIG. 83, and FIG. 85 is a main scanning direction on the digitizer shown in FIG. 78. 86 is a block diagram showing another example of the area signal generation section shown in FIG. 82; FIGS. 87(a) and 87(b) are flowcharts showing other examples of the flowchart in FIG. 80. , FIG. 88 is a block diagram showing still another example of FIG. 82, and FIG. 89 is a flowchart showing another example of FIG. 8I. 210... CCD sensor 403... Feature determination circuit 108... Multiplication coefficient generation unit 115, 114... Multiplier 116... Adder 117... Filter 118... γ conversion--11 ◆Shi Kocho (〃 Agriculture) t (1 KC IC C1K H3YNC No. 13-2 Oka,, /g27 No. 2 difference (4, N No. 7 Figure 1! (17) 2Z!;4 No. 23 -2 Figure Figure 30 Input #ρ/ At the end of Figure 4, 0 SS diagram CLK H, 5yNc- CAM'-1]! 4-1) J -tA αfl) 14 r (Zashi?) L, 4° Fig. 72 (e) Fig. 72 (f) W, 72 Jyo KoS (Note) Setting type 1 in all Oil and water in the above table) Figure 710

Claims (3)

[Claims] (1) Discrimination means for discriminating character parts or halftone parts in the target image; specification means for specifying an image area in the target image; and controlling the operation of the discrimination means with respect to the area specified by the specification means. An image processing apparatus comprising: a control means for controlling an image; (2) The image processing apparatus according to claim (1), wherein the target image is a color image. (3) The image processing apparatus according to claim 1, wherein the control means is means for setting a discrimination condition for the discrimination means with respect to the designated area.