JPH0453350A - Image processing device - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to an image processing apparatus that records color images using recording materials (ink toner, etc.) of multiple colors, such as a color copying machine or a color copying machine. .

[Conventional technology]

When performing color recording by additive color mixing based on color-separated image signals, black thin lines and characters are generally Y, M, and C.
Alternatively, recording materials of four colors Y, M, and C are superimposed and recorded.

[Problem to be solved by the invention]

However, when layering recording materials of 3 to 4 colors, ■ The overlapping colors do not appear completely black; ■ It is difficult to overlap with high precision, so it does not appear black; ■ Based on pseudo halftone processing (in recording There are problems such as the inability to record high-resolution images because dots cannot be recorded in a continuous line.For monochrome images, it is difficult to identify and separate thin lines and character parts from other halftone image parts. The applicant has proposed a method of expressing both the gradation and resolution of an image by recording only the former at high resolution.

However, although this method solves the above-mentioned problem (2), it has the disadvantage that it requires separate hardware for each color when applied to color images, and it also needs to identify black parts from color signals, e.g. The method of determining a region that matches a character part extracted from a single-color G component as a black character region has problems, such as that it is difficult to identify the black part inexpensively and accurately, and that it is difficult to accurately match the two-headed areas. leave. Also, the so-called black signal generated by masking and UCR is used to correct only the black that can be expressed by C, M, and Y, and it cannot be used for this type of processing because it does not have a sufficient level especially in the thin line parts of black characters. do not have.

The present invention has been made in view of the above-mentioned problems, and it is an object of the present invention to provide an image processing device that can satisfactorily reproduce color images, particularly fine black lines in color images.

Another object of the present invention is to provide an image processing device with a simplified circuit configuration.

Another object of the present invention is to provide an image processing device suitable for high-speed processing and real-time processing.

[Means and operations for solving the problems] In order to solve the above problems, the image processing apparatus of the present invention includes means for generating a black signal and other recording color component signals from a plurality of color component signals, The image forming apparatus is characterized by comprising means for identifying a thinner line part from a halftone image part on a pixel by pixel basis, and means for controlling the other recorded color signal based on the output signal of the identifying means.

〔Example〕

FIG. 1(a) shows an entire embodiment of the present invention in a block diagram.

In the figure, 1 is a color CCD line sensor that separates almost the same point on a document into three colors, R, G, and B, and quantizes each color into 8 bits. The R, G, and B signals are input to the color processing section 2, and are subjected to so-called color signal processing such as logarithmic conversion, masking, UCR, and γ conversion to produce recording colors Y and M.

Each 8-bit recording signal is obtained on C. The recording signals are subjected to so-called pseudo-halftone processing in the binarization sections 3, 1.3-72.3-3.3-4, respectively. On the other hand, the color separated R
, G, and B signals are generated by the black signal generating section 6, which is a feature of the present invention.
A black signal d having 8 bits and 256 levels with suppressed color tone is generated. The black signal is converted into high-definition binarization by the binarization unit 7. The binarized signal is input to the black character identification unit 8, where it is separated from the continuous halftone part and the halftone image part to distinguish the text and fine line parts. The signals are identified and separated, and the identification result signal is input to the recording signal control section 4.

The control section controls the previously described independently binarized four-color binary numbers based on the identification signal, and drives the recording element 5.

Now, the embodiment of the present invention will be explained in detail in the order of the black signal generation section 6, the binarization section 7, the identification section 8, and the binary data control section 4.

Note that all known techniques can be applied to the color signal processing section 2 and the binarization section 3, so the explanation thereof will be omitted.

<Black signal generation section> The black signal generation section is shown in FIG. 2(a). In the figure, a maximum value detection unit 61 and a minimum value detection unit 62 each compare the levels of RGB signals represented by 8 bits for each pixel, and calculate the maximum value and minimum value of the levels as m a x (
RGB ) min (RG B ). The adder 63 calculates the difference between the same signals max (RGB) - min
(RGB), the multiplier 66 multiplies the result by a predetermined constant α, and the adder 64 adds it to max(RBG).

If the addition result exceeds 8 bit width, it is limited to 255 by a limiter 65 to obtain a black signal d.

FIG. 2(b) shows an explanatory diagram of this process. In this embodiment, the larger the value of each color signal R, GXB, the whiter it is, that is, R=
When G=B=0, it is assumed that the color is black. Therefore, Fig. 2 (b
), the part A shows a thin line with tint, and the part B shows a black line.

d=max (RGB) 10(t [:max (RG
B) -min (RGB)] -■(max (RGB
) is a gray component signal, may (RGB)min (R
BG) is a signal indicating color, α is a color suppression constant) The physical meaning of this process is max (RGB) - min (
If RGB) is considered a color and max (RGB) is a gray component (brightness), then max (RGB) - m
When in (RGB) is large, that is, at a point that has a tint, this value can be multiplied by a constant 6 and added to may (RGB) to convert it to a brighter (whiter) direction. Therefore, as the value of the constant α increases, α is called a color suppression constant, as it is a value representing the degree to which the tint is suppressed to make a point with a tint whiter. That is, by changing the value of α set in the multiplier 66 by a CPU (not shown), the degree of black component detection indicated by d can be changed. In Figure 2(b), α
= 1, but if max (RGB) is large and max
When x (RGB)-min (RGB) is large, the limiter 65 of FIG. 2(a) takes a maximum value of 255 representing a complete white point. Therefore, it is considered that a change in the generated signal d essentially means a change in the black component.

Now, the black signal d is input to the binarization section 7 shown in FIG.

<Binarization Unit> In FIG. 3, for the black signal d, data for one line each is delayed and held using a memory 71, and data for five pixels sequentially delayed by lines is added by an adder 72. Further, an adder 75 adds five additional values obtained by delaying and holding the added values for each pixel by an F/F 73. If the output of the adder is delayed by two lines from the input data, that is, the output position of F/F 73-6, which is the output of memory 71-2 delayed by F/F 73-5.736, is the pixel position of interest.
This is the integrated value for the surrounding 25 pixels, and therefore, by multiplying it by 1/25 in the divider 76, the average value m of the black signal in the vicinity of the pixel of interest can be obtained.

A comparator 79 compares the data d at the pixel position of interest using the average value m as a threshold value to obtain a more precise binary number B.

That is, shall be.

Further, an adder 77 calculates the difference between the average values m and d, an absolute value circuit 78 converts it to an absolute value, and a comparator 74 compares it with a constant δ to obtain a binary code C.

That is, shall be.

The physical meaning of both signals is that the former B is a signal obtained by high-definition binarization of a black signal, and the latter C is a signal obtained by binarizing the amount of level change at the pixel of interest. If B=1 and C=1, it can be determined that the change in density at the pixel of interest is greater than the constant δ and is changing toward black.

In other words, it can be said that there is a high probability that this point is part of a thin character line.

However, there is a possibility that the pixel of interest is a halftone image portion expressed as a halftone dot, so in order to remove the halftone dot, the same 2-bit signal is input to the black character identification unit 8 to identify the character portion.

Black character identification section> Figure 4 shows the black character identification section. First, the input binary number B is line memory 80-1.80-2.80-3.80
-4 is used to sequentially delay and hold one line at a time, and flip-flops 81-0 to 81-9 delay and hold each pixel. Therefore, if the pixel position of interest is the output position of F/F81-4, the binary data B of 8 pixels adjacent to the pixel of interest are the input position of F/F81-2.81-4.81-6 and the F/F81-4 output position. /F81-2.81-6.813.8 knee 5.
Nine adjacent pixel data including the pixel of interest located at the output terminal 81-7 are input to the gate circuit 832 together.

Similarly, the gate circuit 83-1 is supplied with pixel data one line after the pixel of interest, that is, the nine points adjacent to the F/F 812 output position, and the gate circuit 83-4 is supplied with pixel data one line before the pixel of interest, that is, the F/F 812 output position. /F81-6 Nine points of data adjacent to the output position are input. Inside the gate circuit 83 (as will be explained in detail later), each center point is adjacent to the 8
O to 4 for each pixel, depending on whether the point is inverted from its binary level (0 or 1), that is, whether the central point is isolated from the surroundings and has a level of "0" or "1". The value of S
Assign.

The larger the value S, the greater the possibility that the pixel is a halftone dot pixel, and conversely, if it is O, it is more likely that the pixel is part of a character or the like.

This is because a character line is a collection of one-dimensionally continuous dots. However, whether it is a character or not is not necessarily 1.
Since it cannot be determined based on the point, the value S indicating the isolation divided for each pixel is two-dimensionally integrated for determination. That is, multivalued data indicating the degree of isolation is used for determination. First, adder 85-1 adds 3 pixels in the line direction, and the result is added to 6 pixels using F/Fs 84~1~84-6.
If the pixel delay is held and added by the adder 85-2, and if the pixel delayed by 2 lines and 4 pixels with respect to the input image data B is set as the pixel of interest, data for 7 pixels in 3 parts centered on the pixel of interest will be obtained. The added value Pf of S is determined. The feature quantity Pf means a two-dimensional spatial frequency. In other words, the larger the Pf value, the value of binary data B near the pixel of interest is “0”←→゛″1”
This means that there are many inversions, that is, the spatial frequency is high, and that there are many two-dimensionally isolated points.

Therefore, the Pf value is set to a predetermined constant K (approximately 4 to 5) and the comparator 8.
6, and the pixel position of interest, that is, memory 80-5.
．． 2 lines C with 80-6 and 4 pixels further delayed with F/Fs 82-1 to 824, and 2 lines B at the same position.
The 87 output value E=1 obtained by ANDing is a part of the black character.

The same signal E and one line from the pixel position of interest and F7F891 ~
The signal B delayed by two pixels in step 2 is input to the black character signal generating section 88 to generate black character signals KB and KW as final outputs.

Now, the inside of the aforementioned gate circuit 83 will be explained in detail. FIG. 5 is a diagram showing the gate circuit 83, in which as bz C-, ds ex fSg adjacent to the central pixel i
zh is each input/output signal B of the F/F 81 mentioned above. E
XOR game) 831-1.831-2 is a, i, h
It is detected whether the pixel of interest i is reversed in the direction. In other words, if both EX-OR gate outputs are "1", the AND gate 832-1 output is 1, and the target pixel i is aih
isolated in direction. Similarly, EX-OR gate 8313
．． 831-4, AND gate 832-2 is in the cif direction,
EX-OR gate 831-5, 831-6, AND gate 832-3 is in the big direction, EX-OR gate 831-
7.831-8, AND gate 832-4 detects isolation in the die direction. Here, AND gate 833-1.
8332, OR gate 834 indicates that the target pixel i is continuously at the same level "0" or "l" in the big direction or the die direction.
The OR gate 834 output becomes "0" if the case is detected. The signal is connected to AND gates 835-1-8
AND gates 832-1 to 832-4 mentioned above in 35-4
By ANDing the output and adding the output result in an adder 836, S indicating a value from 0 to 4 is obtained.

The physical meaning of the conditions created by the AND gate 833 and the OR gate 834 is that the continuity is a line segment that intersects at right angles on the paper surface (original or recording paper surface), so it is highly likely that they are part of a character.

Therefore, in this case, the characteristic amount Pf is uniquely set to s20 and the feature amount Pf is lowered as a result. It should be noted that the area to be integrated as the Pf value is desirably rectangular, about 7 parts 7, and is wider than this embodiment so that characters can be identified with high precision. However, the block size and block shape are not limited to the above-mentioned example, and can be set as appropriate depending on detection accuracy and the like.

Next, the inside of the black character signal generating section 88 will be explained in detail with reference to FIG. The same processing part is KBl (lb
The purpose is to generate a pixel position signal KW=1 (lbit) signal adjacent to the it) signal and adjacent to the black character with B=0. Input the lbit signals E and B mentioned above, delay the E signal by line, and send the F/Fs 880-1 to 880-
5 further delays each pixel to F/F the pixel position of interest.
If the output position is 880-3, the OR gate 881 obtains "l" as its output when any of the eight pixels adjacent to the pixel of interest has E=1. Therefore, if the B signal is delayed by the F/F 880-7 and adjusted to the target pixel position, then inverted and inputted to the AND gate 882, a KW times signal can be obtained. Same signal KB
The recording signal control unit 4, which is a feature of the present invention and controls the KW multiplied recording signal, will be explained in detail.

<Recording signal control unit> Because the internal processing to generate the black character signals KW and KB described above is delayed by 5 lines and several pixels with respect to the binarized data,
Delay the value data to match the KB and KW signal positions. The memories 40-1 to 40-4 in FIG. 7 constitute a delay circuit for this purpose, and each binary recording signal KB, KW is as follows: When K B = 1, KW = 1 (7), C, M, Y = Control like 0. In other words, the black recording signal is O of the signal 2 and KB=1.
If recording is performed in R and a color recording signal is present at the same position, color recording is stopped and suppressed to avoid color mixture. Also, KW=1
Similarly, signal pixels can be recorded more clearly by suppressing color recording signals and stopping recording dots around black characters.

The relationship between KW and dot recording in KB, 2 is shown in Table 1 below.

Table 1 Recording device 5> Next, the configuration of the recording device 5 shown in FIG. 1(a) will be described.

The present invention is applicable to various recording methods such as inkjet recording, thermal transfer recording, and electrostatic recording.

Here, an example of use in inkjet recording will be described.

FIG. 1O is a perspective view of the periphery of a head of an inkjet recording apparatus using heating elements.

In the figure, a head unit 51 serves as a total of four nozzles 52. That is, the head unit 51 has the black ink discharge nozzle 52, a 1-arrow ink discharge nozzle 52Y, a magenta ink discharge nozzle 52M, and a cyan ink discharge nozzle 52C. 53 is an ink supply tube, 5
4 is a main tank, and four main tanks are provided corresponding to each nozzle.

The configuration of the nozzle 52 will be explained using the cross-sectional view of FIG. 11. 55 is a top plate, 56 is a bottom plate, 57 is a generating element, 58 is an orifice portion, and 59 is an ink.

When voltage is applied to the heating element 57, heat is generated, bubbles are formed around the element 57, and when the voltage application ends, the bubbles contract. Ink near the orifice portion 58 is ejected from the orifice portion 58 as the bubbles form and contract.

This recording head uses thermal energy to cause state changes in the ink, such as film boiling, to generate air bubbles, and these bubbles are used to direct the ink to the ejection ports (nozzles).
This is a so-called bubble jet type recording head that records characters, images, etc. by ejecting the liquid from the liquid toward the recording material. In this recording head, the size of the heating resistor (heater) installed in each nozzle is much smaller than the piezoelectric elements used in conventional inkjet recording, making it possible to have multiple nozzles at a high density. , high-quality recorded images can be obtained, and it has features such as high speed and low noise.

A recording device having multiple heads for one type of coloring material may be used.

In FIG. 12, numeral 101 is a head unit in which a large number of inkjet heads for one type of coloring material are arranged in the sub-scanning direction, and includes units for black, yellow cyan, and magenta. 107 is an ink tank for each head unit, 109 is a signal line, and 104 is a carriage drive motor that moves the carriage 105 to which the head unit is attached along the rail 103 in cooperation with the conveyor belt 105. 106 is a recording paper, 120 is a platen,
Reference numerals 111 and 112 indicate recording paper transport rollers, 113 indicates a recording paper roll, and 114 indicates a guide roller, respectively. The head unit 101 is composed of a plurality of inkjet heads using the heating elements shown in FIG. 11, but it is of course possible to use inkjet heads using electromechanical conversion means such as piezo elements.

<Another embodiment of the black signal generation section (1)> FIG. 2(c) shows another embodiment of black signal generation.

In the previous example, the tint suppression constant is the gray component max (RGB)
Although the value α is constant regardless of the value, the gray component is generally large,
In other words, the brighter the image, the more likely it is to have a tint, so the degree of tint suppression is increased as the brightness increases. By doing so, a black signal d with higher contrast can be generated.

d = max (RBG) max (RGB) −min obtained by adder 63
(RGB) is multiplied by the output max (RGB) of the maximum value detection unit 61 in a multiplier 661, and the result is divided by a constant β in a divider 662. A good value of β is, for example, about 128. Further, the adder 64 adds the result to max (RGB), and the limiter 65 clamps it to 255.

Another embodiment of the black signal generation section (2-)> In the previous embodiment, the color tone was defined by the difference between max (RGB) and min (RGB), and the suppression was performed by addition operation, but if the color tone is (R
GB) and m1n (RGB), and the suppression is defined as max(
RGB) may be defined by a multiplication operation. In other words, the constant γ is preferably about 63 here. According to the above equation, the brighter the gray component is, the more effective the suppression of color tone is, and the degree of suppression is determined by the constant γ, that is, the region m
If both ax (RGB) and m1n (RGB) are small, it will be relaxed, so even if the black component has some tint, the value of d will be thicker (rather than the black signal d with higher contrast). Can be generated.

In the embodiment shown in FIG. 2(d), an adder 663 adds a constant γ to the output of the minimum value detector 62, and a divider 664 divides the output of the maximum value detector 61. Then, in the multiplier 665, m a
Multiply by x(RGB).

As described above, the algorithm for generating multi-value data with suppressed color from R, G, and H data is not limited to the present invention, and is not particularly limited to primary colors RGB, but can also be used for recording colors such as YMC. It goes without saying that similar effects can be obtained by using other color component signals such as .

<Another embodiment (1) of recording data control unit> In the above-mentioned embodiment, an example was disclosed in which recording data is controlled after binarization.
An example of controlling multivalued data before binarization is shown using FIG. 1(b) and FIG. 8.

In FIG. 1(b), C, M, and Y are processed by the color signal processing section.

The data converted into the K4 color signal is controlled by the 2-bit signals K W and K 11 output from the black character signal generating section 8, and then binarized by the binarizing section 3 and recorded. In the case of this embodiment, if the recording element 5 is an LBP in which the dot size can be modulated in multiple steps by pulse width modulation, or a thermal recording element capable of recording dots in shading, such as a piezo type ink jet printer or a thermal transfer printer, the recording element 5 is binary. The conversion unit 3 is not necessarily necessary, and the controlled multi-valued data may be used for recording as is.

Now, the inside of the recording signal control section will be explained in detail using FIG. 8. In FIG. 8(a), the memory 91.92°93,9
4 constitutes a delay circuit for aligning with the black character signals KW and KB as described above, and directly sends multi-value data to the LUT of ROM 95, 96, 97.98 according to KB and KW for each pixel of each color. Convert.

In the case of a black signal, when KB=1, the input data is emphasized toward black as shown in Figure 8 (C), and the color signals C, M, and Y are reversed as shown in Figure 8 (B). A γ table that emphasizes the white direction is selected and used.

Further, when KW=1, the degree of emphasis of the color signals C, M, and Y signals is slightly relaxed compared to when KBl is used. In this embodiment, in order to control the multi-value level, especially when K W = 1, the periphery of black characters having a background color can be controlled to a naturally light color, so that a recorded image of higher quality can be obtained.

It goes without saying that in the case of KB=1, CK=255, qM, CC, CY=0 may be set regardless of the input data.

<Another embodiment (2) of the recording data control unit> In the above embodiment, black character processing was performed after masking UCR, but the first
As shown in Figure (c), a similar effect can be obtained by directly controlling the primary color R, G, and B signals. The recording data control unit IO in FIG. 1(C) will be explained in detail using FIG. 9.

In FIG. 9(a), a delay memory 1 similar to the above embodiment is shown.
01. 102. 103 is a memory having the same capacity as 91.92.93.94 in the embodiment shown in FIG.
In this embodiment, since memory for one color is not required, it can be realized with a simpler configuration. Now, data conversion is done in ROM1 for each color.
This is performed using LUT conversion using 04.105.106.

As shown in Fig. 8(C), when K W ==: 1, each color is emphasized towards white, and conversely, as shown in Fig. 8(d), when KB =
If it is 1, it will be emphasized towards black. By applying normal masking and UCR to the converted data, if KW=1, Y
, M, C, and K are all emphasized toward white, and conversely, KB=1
, the signal is emphasized toward black, while the C9M and Y signals are emphasized toward white. In addition, in the above embodiments, the conversion characteristics are not limited to linear characteristics, but if nonlinear characteristics are provided, K
When B=1, the input signal can be controlled to be blacker (and when KW=1, the color signal can be controlled to be whiter).

As described above, according to the above embodiment, black characters and thin lines in images can be identified at low cost and with high precision;
Therefore, it is possible to record text in a single black color with high resolution, and (1) even if color shift occurs during recording, it is possible to record black text clearly in black.

■In particular, according to this embodiment, since the black signal is extracted and the thin line portion and the halftone image portion are identified based on the black signal, there is no need to generate a new black signal for identification, and the circuit configuration can be easily done.

Note that as a printer for outputting the recording signal generated as described above, a printer capable of color recording such as a color inkjet printer, a color thermal transfer printer, a color tod printer, a color laser beam printer, etc. can be used.

It is particularly effective for printers using a type of head that ejects droplets using film boiling caused by thermal energy, as shown in US Pat. No. 4,723,129 and US Pat. No. 4,740,793.

Further, the above-described arithmetic circuit may be configured using ROM, RAM, etc., or may be realized by computer software that can perform the above-mentioned arithmetic operations.

Further, image data may be input from a CCD sensor, from a host computer via an interface, or from memory.

〔Effect of the invention〕

As described above, according to the present invention, fine black lines in a color image can be reproduced satisfactorily.

[Brief explanation of drawings]

1 is a diagram showing the overall configuration of an embodiment of the present invention, FIG. 2 is a diagram showing black signal generation, FIG. 3 is a diagram showing a binarization section, FIG. 4 is a diagram showing a black character identification section, FIG. 5 is a diagram showing a halftone dot determination circuit, FIG. 6 is a diagram showing a black character signal generation section, FIG. 7 is a diagram showing a recording signal control section, and FIGS. 8 and 9 are diagrams showing another example of the present invention. FIGS. 10, 11, and 12 are diagrams illustrating the recording apparatus. 6... Black signal generation section 8... Black character identification section KW KB qth KB (W KW KB (b) CC) (-, T, M in qth (b) (C) (〆 2CrB

Claims (1)

[Claims] means for generating a black signal and other recorded color component signals from a plurality of color component signals; means for identifying thin line areas from the black signal from halftone image areas for each pixel; An image processing device comprising: means for controlling recording color signals.