JP2965890B2 - Image processing device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing apparatus for correcting recording unevenness or display unevenness.

[0002]

2. Description of the Related Art A plurality of recording (or display) elements (hereinafter, referred to as "recording or displaying" elements)
“Record (or display)” is simply expressed as “record” and includes display. When a recording element is used in a printer, it may be referred to as a “recording head”. In the case of printing an image using the method (1), if there is uneven density due to each recording element, the image quality is deteriorated due to the uneven density when reproducing a halftone image.

Japanese Patent Application Laid-Open No. 4-8571 is known as a technique for correcting and correcting the density unevenness of the printing element as described above. This technology reads the density of a test pattern recorded by a multi-nozzle jet recording apparatus, finds the average density, multiplies the reciprocal of the ratio of the density of the test pattern of each recording element to this average density as a correction value, Perform unevenness correction. In the case of color, since the density characteristics are different for each type (color) of the recording head, the correction reference level is changed according to the type of the recording head. However, in this case, it is necessary to suppress the density unevenness of each recording element to several percent or less in order to avoid deterioration of image quality, and there is a problem that the density control of each element requires extremely high accuracy. . In other words, there is a problem that the density unevenness in the recording head requires higher performance than the gradation accuracy required for the original display, and it is very difficult to correct the unevenness.

When a recording sheet is fed in the sub-scanning direction by recording elements arranged in a line and two-dimensional halftones are recorded, for example, a 300 dpi recording head differs slightly depending on the noise level at the time of recording. Regarding the density unevenness that is worrisome at a distance, even if the density unevenness is several percent (1-2%), streak-like noise becomes conspicuous and hard to see, so it is necessary to suppress the density unevenness to several percent or less. In order to suppress density unevenness to several percent or less, it is necessary to perform recording control at a control level of about 100 gradations in terms of the number of gradations.

However, in the reproduction of a natural image, almost sufficient reproduction without a sense of incongruity is possible even at about 64 gradations, and in some cases, 32 gradations can be satisfied. When a technique such as multi-level error diffusion recording or a dither method is used, it is possible to reproduce a natural image by recording with a smaller number of gradations. Therefore, higher accuracy (control level) is required to correct the unevenness of the recording elements than to correct the number of tones required for reproducing a halftone of a natural image. Since the recording density unevenness due to a line head or the like becomes fixed and streaked regardless of the image, it can be said that the noise is extremely noticeable. Therefore, there is a problem that performance that is more accurate than the gradation accuracy required for the original display is required only because of the density unevenness of the recording element.

A technique for correcting unevenness with high accuracy using a plurality of pixels instead of one pixel is disclosed in Japanese Patent Laid-Open No. 4-339364.
It is described in. It is known that this technique corrects density unevenness of m gradations by a plurality of k pixels in a printing apparatus capable of outputting halftones of n gradations with one pixel. In this case, k
Is large, high-precision correction is possible even if the gradation control of one pixel is small. However, this technique has a disadvantage that pixels for displaying an image are effectively coarse.
Further, when the density unevenness is finely changed, it is not possible to cope with such density correction using a plurality of pixels. In general, in the case of a line formed by a solid-state head using an LED element or an ink jet, the recording element may change for each dot, so that the density may change very finely.

According to Japanese Patent Application Laid-Open No. 4-217172,
When recording an image signal in a pseudo halftone expression similar to the error diffusion recording method, binarizing (quantizing) the image signal based on data of the recording element unevenness when binarizing (quantizing) the image signal. It is known to correct the unevenness of the recording element by using. This method enables more accurate correction of recording unevenness by calculating an average threshold value from a plurality of data in the vicinity of a pixel of interest and binarizing the data at the time of this binarization (quantization). . The density is corrected by diffusing an error in the binarization (quantization) at this time. This method has the following features. (1) The recording density unevenness can be corrected even with a binary printer or a quantized multi-value printer. That is, since similar characteristics to error diffusion recording are used, correction can be performed with a smaller number of gradations than the number of gradations required for unevenness correction. (2) By quantizing the data of the target pixel from the average value in consideration of the weight of the peripheral unevenness data of the target pixel, more accurate unevenness correction can be performed.

However, in this method, the response is reduced with respect to a steep unevenness having a high-frequency component (ie, a steep change in the characteristics of adjacent elements in the element array). Further, since this method has a characteristic that an error depending on an error of density unevenness is propagated to peripheral pixels, such an unevenness that changes sharply for each pixel as seen in an exposure recording apparatus including an LED array or the like is obtained. With respect to a recording head having unevenness in the high-frequency component, if such correction is made, the correction error extends to the peripheral pixels, causing fluctuations in the low-frequency component. Therefore, such correction further degrades the image quality before the correction is performed. In addition, since this correction method corrects fixed noise caused by the recording head, the correction error that could not be corrected by one pixel is diffused to the periphery, resulting in large noise in area, so that image quality is significantly deteriorated. However, it has a drawback that it is difficult to correct the density unevenness of the recording element.

[0009] As a technique similar to the above technique, 1
When a pixel is represented by multi-value using a plurality of different printing elements, the unevenness is reduced by using the plurality of printing elements, and at the same time, the quantization (multi-valued) is performed in consideration of the unevenness characteristic of each of the plurality of printing elements. (Japanese Patent Application Laid-Open No. HEI 5-183738).

However, this technique also basically determines the threshold value of the pixel of interest based on the periphery of the pixel of interest and quantizes it, similarly to the above-mentioned Japanese Patent Application Laid-Open No. 4-217172, so that a plurality of recording elements are obtained. Are quantized in consideration of the non-uniformity characteristics of each of them, thereby improving the non-uniformity correction accuracy. However, this technique also has a high-frequency component, and the response to the unevenness that has changed sharply deteriorates, as in the case of Japanese Patent Application Laid-Open No. Hei 4-217172. It has disadvantages such as deterioration.

Further, when a uniform image is recorded by the error diffusion processing, a pattern unique to the error diffusion, that is, a texture is generated, and the image quality may be deteriorated. Therefore, it is known to add a periodic pattern for the purpose of removing the inherent noise of error diffusion (“Binary image processing ASIC selective emphasis error diffusion method”, Journal of the Institute of Image Electronics Engineers of Japan 1991 VOL.20 No.5p43)
6-449 and "Binarized representation of mixed images of characters / dots / photographs-Continuous adaptive binarization using image area separation variables-" Journal of the Institute of Image Electronics Engineers of Japan 1991 VOL. 20 No. 5 p476-483). The method of adding the periodic pattern to the input signal is not preferable because the effect of the additional signal appears on the output signal as it is. That is, depending on the additional pattern, a unique texture is conspicuous and causes deterioration in image quality.

[0012]

In general, when correcting unevenness that can be seen by one pixel with one pixel, such density unevenness correction using a plurality of printing elements (array printing elements) is performed every time one pixel is printed on an input image signal. If the density unevenness is corrected and recorded, the unevenness can be corrected. However, since the accuracy needs to be corrected to the limit of the visual acuity as described above, the accuracy needs to be corrected to several percent or less. As the recording signal, the input image signal is signal-corrected to several percent or less each time one pixel is recorded. However, at the same time, the recording apparatus is also required to perform recording with an accuracy of several percent or less. In this case, it is required to correct the density unevenness with higher accuracy than the multi-valued number required for the halftone representation of the image. Further, in the method of correcting density unevenness by using a method similar to error diffusion recording, correction is not performed in units of one pixel, but processing is performed on a plurality of pixels, and an error depending on the characteristic of the unevenness is also diffused. This makes it difficult to correct uneven density including high-frequency components.

An object of the present invention is to make it possible to perform sufficient unevenness correction even for steep unevenness including a high-frequency component that changes for each pixel with the minimum number of tones required to represent an image. In addition, an object of the present invention is to provide an image forming apparatus capable of correcting unevenness inherent in error diffusion.

[0014]

In order to solve the above-mentioned problems, the present invention quantizes an image signal, propagates the quantization error to peripheral pixels, performs error diffusion, and records / quantizes the quantized signal. In an image processing apparatus for recording or displaying by inputting to a display element, adding means for adding an input image signal and an error signal depending on the quantization error, and recording / displaying the output signal of the adding means in the recording / display Correction is made in accordance with density unevenness data indicating density unevenness of recording or display by the element, and the signal after this correction is quantized to obtain the recording / display element.
Estimating the unevenness correction quantization means for supplying to the child, the density of the recorded or displayed by the recording / display device based signal quantized by the non-uniformity correction quantization means and the density unevenness data, corresponding to the concentration concentration to output the concentrations signal
Estimating means, error calculating means for calculating an error between a density signal output from the density estimating means and an output signal from the adding means, and an error calculating means for calculating an error based on the error calculated by the error calculating means.
There are characterized by comprising a means for generating said error signal.

Further, the present invention quantizes an image signal, propagates the quantization error to peripheral pixels to perform error diffusion, and inputs the quantized signal to a recording / display element to record or display an image. in the processing unit, and adding means for adding the error signal depending on the image signal input to the quantization error, quantization levels than the output signal of said adding means
A signal adding means for adding a pattern signal consisting of a periodic pattern or a random pattern having a lower amplitude, and recording by the recording / display element an output signal of the adding means after the pattern signal is added by the signal adding means. Alternatively, correction is performed in accordance with density unevenness data indicating density unevenness of display, and the corrected signal is quantized to perform the recording / display.
Estimating the unevenness correction quantization means for supplying to the device, the density of the recorded or displayed by the recording / display device based signal quantized by the non-uniformity correction quantization means and the density unevenness data, corresponding to the concentration concentrated to output the density signal
A degree estimating means, wherein the error calculation means for calculating an error between the output signal of the density signal and said adding means output from the density estimating means, based on said calculated by the error calculating unit error
Zui and characterized by comprising a means for generating said error signal.

In the present invention, the unevenness correction quantizing means, the density estimating means and the error calculating means may be constituted by a single ROM as a table. Further, in the present invention, an image recorded or displayed may be read when a predetermined signal is given to the recording / display element, and the density unevenness of the image may be used as the density unevenness data.

In the present invention, the input image signal depends on the quantization error .
The signal obtained by adding the existing error signal according to the uneven density data.
Density unevenness correction to correct this density unevenness.
In order to perform quantization on the positive signal, the quantization process
Concentration unevenness higher than the quantization level existing on the floor
It is corrected up to the activation level. Quantization is used for recording / display elements.
The number of gradations required for control, for example, a resolution of 300 dpi
In this case, it is performed at about 8 to 16 gradations. At this time, the quantized signal
No. is corrected according to the density unevenness of the recording / display element
Therefore, recording / display elements with thin recording or display
Is a record / table where large values are recorded or displayed
When small values are input to the indicator elements,
Will be corrected. To correct uneven density, specifically,
The reciprocal of the test pattern density divided by its average is
This is performed by multiplying the addition signal.

Here, the quantization level is generally 8 to 16 floors.
Tone, and to correct unevenness of several percent below the quantization level
Insufficient density is added to the quantized signal.
The density at the time of actual recording / display of taste is determined, and
Density signal and the value before density unevenness correction, that is, input
Add error signal depending on quantization error to image signal
Calculate the error with the signal that
Multiplication by the weighting factor
A difference signal is being generated. In this way, it is calculated
The error is not a correction error of the density unevenness of the recording / display element,
Reflected in the quantization error, the error diffusion
This is almost the same as the quantization error when there is no deviation, and the fixed unevenness information depending on the recording head is extremely small.
Therefore, even if there is a sharp change in the density unevenness,
The correction error is reduced to the quantization level given to the recording element for one pixel.

In general, the optimum quantization level in a recording / display element is determined from a visual limit curve (limit of disappearance), and it is appropriate to determine the quantization level to a level that cannot be visually identified. In other words, the fluctuation of the quantization level is an invisible level. Therefore, the density unevenness of the recording / display element is less than the quantization level, so that it becomes invisible and sufficient unevenness correction can be performed.

In the case of a uniform density output such as a CG output, the output from the array recording element becomes streak-like noise. The effective spatial frequency in the scanning direction decreases, and noise near the eyesight limit may be noticeable.
However, in such a case, the input image signal is quantized.
A periodic pattern is added to the signal obtained by adding the error signal depending on the error.
Or add a pattern signal consisting of a random pattern
By adding a periodic variation of less quantization levels Te, relaxed quantization levels below the noise due to density unevenness of printing elements, can also be suppressed further error diffusion inherent noise.

[0021]

Embodiments of the present invention will be described with reference to the drawings. (First Embodiment) FIG. 1 shows a block diagram of a first embodiment of the present invention. In FIG. 1, a multi-value printer 101 is a printing apparatus including an array printing element. More specifically, it is a recording device including an electrophotographic printer including exposure using a light emitting diode array (LED). In this device, 16 gradations can be obtained by controlling the pulse width of the light emission time. However, there is about 20% density unevenness due to light emission variation of the LED array. For example, FIG. 2 shows a part of the density unevenness in the case where the exposure is performed with an element having uneven light emission (that is, an element array having uneven light emission between the elements) and the recorded density is measured. In FIG. 2, the vertical axis indicates the recording density corresponding to the brightness of the light emitting element, and the horizontal axis indicates the position of the light emitting element. In FIG. 2, the density unevenness is 1 on average.
It is about 0%, but at most 25% or less. In the apparatus of the present embodiment, the emission intensity measured by emitting light for each pixel is recorded in the density unevenness data recording ROM 102 as unevenness correction data. In the first embodiment shown in FIG. 2, since the maximum density unevenness is 25% or less,
As the unevenness correction data, the lower 6 bits quantized by 8 bits may be recorded in the density unevenness data recording ROM 102.

8 bits of the image signal are input from the input unit 103 and input to the adder 104. This adder 104
Has a function of adding an error that propagates as an error during multi-value recording, which will be described later. The output of the adder 104 is input to the unevenness correction quantization unit 105.

The unevenness correction quantization unit 105 calculates a value obtained by dividing the density unevenness data of the recording element to the output of the adder 104 performs the unevenness correction, the signal after the non-uniformity correction 4 bits (16 levels) And output. That is, L
When the density becomes low due to ED light emission variation, control is performed to increase the density. In addition, when the density is recorded high, control is performed so that the recording is performed slightly. At this time,
The quantization size is set so that the maximum recording error is the quantization error. The structure of the specific circuit of the unevenness correcting quantization unit 105, an image input unit 8 bits, with density unevenness correction data input unit 6 bits in total 14 bits input, ROM or outputs 4 bits as unevenness correction quantization output R
AM may be used . The 4-bit output is supplied to the multi-value printer 101 and printed. At the same time, the 4-bit output data is input to the unevenness correction density estimating unit 106 together with the 6 bits output from the density unevenness data recording ROM 102.

The unevenness correction density estimating unit 106, based on the 4-bit output data is recorded control signal, estimates the density value which would multilevel printer 101 records. Concentration of example <br/> Baal image point if the normal signal is not obtained only 1/2 of the concentration, from non-uniformity correction quantization unit 105 outputs a control signal of twice the strength. Based on this signal, the multi-valued printer 101 performs printing with a signal having twice the intensity, so that the image point is printed at almost normal density.

Further, when the output of the adder 104 is divided by the density unevenness caused by the recording element, the unevenness correction quantization unit 105 does not perform the uneven density correction unless the quotient is set so as not to overflow. For example, if there is a density decrease of 0.833, the answer of the result of division is 1.2, and if this value is the maximum value of the quotient when divided by the density unevenness of this element, the value is taken as the printer value. It is necessary to correspond to a multi-valued number (for example, four values) of the section. This means that when the input image is 8 bits, if there is no density unevenness, the image quality is directly obtained by the four-valued error diffusion method, but if the value divided by the density unevenness is 1.2, the quantization corresponding to 20% is obtained. It means that the number is reduced. That is, an image quality of a quantization level corresponding to 4 / 1.2 = 3.33 value is obtained. Therefore, when there is uneven density, it is preferable to increase the quantum number of recording or display by that much. This can be considered that the overall image quality is reduced by the amount by which the local density unevenness is corrected by the method of the present invention using the error diffusion method.

Outputs from the adder 104 and the unevenness correction density estimating unit 106 are input to an error calculating unit 107. The error calculating unit 107 outputs a signal actually output to the multi-value printer 101 and an unevenness correcting quantization. An error from an input signal to the unit 105 is calculated. The calculation error is temporarily stored in the error buffer 108, and the calculation error is output from the error buffer 108 at a predetermined timing.

The calculation error output from the error buffer 108 is multiplied by the weighting factor stored in the error diffusion weighting factor unit 109 by a multiplier 110, and the result is multiplied by the image signal from the input unit 103. Added at 104,
Error diffusion is performed.

The additional pattern 112 is a pattern for adding a regular periodic pattern for performing correction other than, for example, density unevenness. In general, when a uniform pattern is input in the error diffusion process, a unique pattern is generated, and particularly at a low density, a visually unsightly pattern may be generated. In such a case, if an additional pattern 112 such as a pattern having a proper cycle, for example, a 4-pixel cycle in both the main scanning and the sub-scanning is added, the change of the micro recording pattern is affected by the additional pattern 112. The occurrence of coarse noise inherent to error diffusion is eliminated.

The operation of the first embodiment will be described with reference to FIG. FIG. 3 is a diagram showing a signal flow in the first embodiment of the present invention. In FIG. 3, f is an input signal,
m is data of uneven density, and Dn is an error diffused from peripheral pixels. In addition, Q is shows the quantization function.

As is apparent from FIG. 3, the multi-value printer 1
As the output g, which is the control signal of 01, it is necessary to give to the multi-valued printer 101 a control signal that is weaker as the density is higher for the same signal strength. Therefore, the quantization is
When the function Q is a quantization function, the function is performed so as to be a function proportional to the reciprocal of the density unevenness m, that is, Q ((f + Dn) / m). The control signal is a function of the position in the main scanning direction.

The actual recording density estimated from the output g of the multi-value printer 101 is the density obtained by multiplying the control signal by the density unevenness m, that is, Q ((f + Dn) / m) × m.
The difference e between the density actually expected to be recorded and the density signal f + Dnl of the input plus the error diffusion before the error signal is obtained by e = f + Dnl-Q ((f + Dn) / m) × m
Becomes Dn1 is an error diffusion signal from the immediately preceding pixel.
Show. Therefore, the error e propagated by the slowly changing pixel
Is less than the quantization error.

Further, propagates recording density error based on the quantization error and the control signal, the error diffusion loop to the input signal f and the output signal g is coincident (addition → quantization → error diffusion → addition of loops ) Works. Further, the propagation of the error with respect to the steep fluctuation mainly depends on the input signal f, and the density unevenness m
Hardly depends on Therefore, even if there is a sharp change in the density unevenness, the amplitude does not become conspicuous, and the amplitude immediately becomes smaller than the quantization error. In general, quantization is set to a level at which quantization is inconspicuous by visual observation, so that by performing such processing, density unevenness generated by the recording element can be corrected.

FIG. 4 is a diagram showing an example of a conventional configuration for further clarifying the effect of the first embodiment of the present invention. The configuration of FIG. 4 differs from the present invention in that the input signal f is immediately divided by the density unevenness m to perform density unevenness correction. In FIG. 4, the symbols used are the same as those in FIG.

More specifically, in FIG. 4, the signal input to the error diffusion loop (addition → quantization → error diffusion → addition loop) is f / m. Assuming that the diffusion error is Dn, the output g is Q (f / m + Dn). Therefore, as the error signal diffusion, the difference e between the control signal Q corresponding to the density estimated to be actually recorded and the sum f / m + Dnl of the input and the density signal of the previous error diffusion is f / m + Dnl− Q
(F / m + Dn). Also in this case, the error e propagated by the pixel which fluctuates slowly becomes equal to or smaller than the quantization error. However, if there is sharp density unevenness, for example, if this division is performed on the order of 8 bits, the density unevenness can be corrected because it is larger than the recording quantization. May not be able to be corrected immediately, in which case the error propagates, and a steep change in density m is converted into a change in low frequency to correct, and the average density is corrected. , Uneven density may be noticeable.

The unevenness correction quantization unit 105, the unevenness correction density estimating unit 106, and the error calculation unit 107 in FIG. 1 may be constituted by one ROM, and this may be used as the unevenness correction quantization error table 111. The input of this unevenness correction quantization error table 111 is image input 8 in this embodiment.
It is possible to configure a total of 14 bits of 6 bits and density unevenness data and a total of 8 bits of output including a control signal of 4 bits and error diffusion data of 4 bits. With such a configuration,
The configuration becomes easy.

In this embodiment, an electrophotographic printer using an LED exposure head has been described. However, other recording systems can be applied. That is, the present invention is also applicable to ink jet recording, thermal transfer recording using a thermal head, and a sublimation printer. The present invention is suitable for correcting density unevenness in a recording system using a head in which a plurality of pixels are arranged in a line. If the density unevenness is fixed, it can be applied to a laser printer.

Although the description has been made with reference to a monochrome printer, in a color printer having different color recording heads, the processing used in this description may be performed independently for each color. When a recording head common to each color is used, the processing may be performed with the same density unevenness data in common for each color.

Further, in this embodiment, an example in which the present invention is applied to a printer is described. However, the present invention is not necessarily limited to a printer, but can be applied to, for example, a liquid crystal display. In a liquid crystal display, depending on the drive control method, a certain density unevenness may be generated vertically or horizontally. In such a case, the density unevenness can be processed with one-dimensional correction data as in the present embodiment. However, generally 2
Since the density unevenness is considered to be dimensional, in this case, the correction may be performed by expanding the density unevenness data to two dimensions.

(Second Embodiment) Since the correction of density unevenness in the correction processing according to the first embodiment is equal to or lower than the recording quantization level, the relation between the resolution and the number of gradations in recording is limited by the visual acuity limit. If it is set as described above, the density unevenness becomes noise below the visual acuity limit, so that it should not be visible. However, when a pattern having a uniform density is output, unevenness due to the print head is corrected. Therefore, the actual print pattern is slightly changed so as to eliminate the density unevenness corresponding to the print head. Since this recording pattern is substantially constant in the sub-scanning direction, the recording pattern may be conspicuous when the recording sample is held diagonally. In such a case, it is preferable to add the additional pattern 112 as shown in FIG. In the first embodiment, the additional pattern 112 is preferably added between the point where branching is performed to perform the error calculation unit 107 and the unevenness correction quantization unit 105. The additional pattern 112 is a periodic pattern of about several pixels, and its amplitude is preferably equal to or less than the quantization level of the unevenness correction quantization unit 105.

When the above-described periodic pattern is added before quantization, the quantization varies depending on the added pattern, and a fine recording pattern changes. For example, when there is no additional pattern 112 for a uniform input signal, there is only a change in the recording pattern depending on the uneven density data. However, when such an additional pattern 112 exists, both the uneven density and the additional pattern 112 are present. Changes in the recording pattern depending on. However, since the effect of the additional pattern 112 does not depend on the error diffusion signal e, the average density does not change even if the additional pattern 112 having a large amplitude is added. Only a local change in the recording pattern (a recording pattern of several pixels) occurs. Therefore, unlike a recording pattern that depends only on density unevenness, the pattern is not always constant and deviated in the sub-scanning direction, and is almost inconspicuous because it is a uniform pattern as a whole.

The additional pattern 112 is stored in the error calculator 1
07 and the point at which the unevenness correction quantization unit 10 is used.
5 and a signal added for the purpose of removing the inherent noise of error diffusion (“binary image processing A
SIC Selective Enhancement Error Diffusion Method "Journal of the Institute of Image Electronics Engineers of Japan 1991
VOL. No. 5, p. 436-449) is not preferable because the effect of the additional signal is the output signal as it is in the method of adding it to the input signal. That is, depending on the additional pattern 112, the additional pattern 112 (specific texture) may be conspicuous and the image quality may be degraded.

In the second embodiment, the solid-state recording head is used in combination with the unevenness correction processing circuit. However, it is sufficiently effective to use it alone. FIG. 5 is a circuit processing block diagram in which an additional pattern 112 is added to a block diagram of a normal error diffusion process. As described above, the additional pattern 112 locally affects the quantization processing unit 501 to cause a change in the pattern of the recording pixel. However, since the additional signal is not added to the input signal, the average somewhat macro density is preserved and no change occurs. That is, there is an effect of controlling only a change in a micro recording pattern. In general, when a uniform pattern is input in the error diffusion process, a unique pattern is generated, and particularly at a low density, a visually unsightly pattern may be generated. In such a case, if an additional pattern 112 such as a pattern having a proper cycle, for example, a 4-pixel cycle in both the main scanning and the sub-scanning is added, the change of the micro recording pattern is affected by the additional pattern 112. The occurrence of coarse noise inherent to error diffusion is eliminated.

The additional pattern 112 is not necessarily a periodic pattern, but may be a random pattern generated by the generation of random numbers. Further, a high frequency component of the input image may be added as the additional pattern 112. In this embodiment, since the correction is not for noise (inherent texture) which is constant in the sub-scanning direction as in the solid head, the present invention can be applied to any recording system, and can be applied to a laser recording device and a liquid crystal display device. is there.

(Third Embodiment) In a third embodiment, an embodiment of a system having an image input device will be described. In the third embodiment, a copier, a facsimile, or an apparatus in which a scanner and a printer are connected can be considered. Here, the copier will be mainly described.

First, the recording unit 601 outputs a test chart which is a uniform gradation recording pattern. The density unevenness of this test chart is measured by the reading unit 602.
Note that this reading unit 602 is similar to a normal scanner.
The reading unevenness is corrected by the reference pattern. Therefore, the unevenness of the read signal is recorded by the recording unit 60.
1 causes unevenness.

In order to avoid the influence of sudden noise at the time of recording or noise such as dust at the time of reading, a plurality of lines are read and an averaging process is performed to reduce the influence of the noise. Thus, the density unevenness data m is obtained. By sending the density unevenness m to the printer unit 601 and performing the processing of the first embodiment, density unevenness correction can be performed.

Actually, the test chart reading section 60
Depending on the position set to 2, there is a possibility that a position shift occurs between the pixel to be corrected and the data. If the position is displaced, it is not possible to correct the unevenness, but rather the unevenness may be amplified. Therefore, in order to confirm the correct position, the density unevenness data is moved and corrected by one pixel at a time, and each data is divided and recorded on one sheet of recording paper. That is, as shown in FIG. 7, correction is performed by moving one pixel at a time in the order of 701, 702, and 703. In the recording by the line head, since it is constant in the sub scanning direction, it moves ± (several pixels) in the main scanning direction,
By fixing at the position where the correction is best performed, it is possible to correct the positional deviation. These processes are performed by the processing unit 603. The present invention is not limited to the above embodiments of the present invention, and it is needless to say that various modifications can be made without departing from the spirit of the present invention.

[0048]

According to the present invention, the following effects can be obtained. In the present invention, a signal that propagates as an error signal in the recording processing system does not propagate a conventional correction error of density unevenness but propagates almost a quantization error of recording control. That is, the quantization error is substantially the same as the quantization error when recording is performed by a recording head having no density unevenness. Therefore, fixed unevenness information depending on the recording head becomes extremely small. Further, even if the recorded output has a sharp change in the characteristics of the recording element, the error due to the correction is reduced to the quantization level given to the recording element in one pixel. That is, since the level becomes lower than the level that cannot be visually identified, the unevenness disappears, and the correction can be sufficiently performed. Further, even if the output has a uniform density such as a CG output, by adding a periodic variation below the quantization level, noise below the quantization level due to the density unevenness of the recording element is reduced, and furthermore, the error diffusion inherent in the error diffusion is further reduced. Noise can also be corrected.

[Brief description of the drawings]

FIG. 1 is a diagram illustrating a density unevenness correction processing block according to a first embodiment.

FIG. 2 is a view showing the non-uniformity characteristics of a write-type solid head used in the first embodiment.

FIG. 3 is a simplified processing diagram for explaining effects of the first embodiment.

FIG. 4 is a simplified processing diagram for explaining a conventional concept.

FIG. 5 is an exemplary view for explaining processing blocks for reducing unique pattern noise in the error diffusion method according to the second embodiment;

FIG. 6 is a view for explaining density unevenness correction processing blocks in a system having a scanner unit and a printer unit.

FIG. 7 is a view for explaining an output method for correcting a displacement of correction data.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 101 ... Multi-value printer 102 ... Density unevenness data recording ROM 103 ... Input unit 104 ... Adder 105 ... Unevenness correction quantization unit 106 ... Unevenness correction density estimation unit 107 ... Error calculation unit 108 ... Error buffer 109 ... Error diffusion weighting coefficient Unit 110: Error diffusion coefficient integration calculation unit 111: Uneven correction quantization error table 112: Additional pattern 501: Quantization processing unit 601 ... Recording unit 602 ... Reading unit 603 ... Processing unit

────────────────────────────────────────────────── ─── Continuation of the front page (56) References JP-A-3-274965 (JP, A) JP-A-4-217172 (JP, A) JP-A-4-207462 (JP, A) JP-A-4-207 351068 (JP, A) JP-A-5-114999 (JP, A) (58) Fields investigated (Int. Cl. ⁶ , DB name) H04N 1/40-1/409 H04N 1/46 H04N 1/60

Claims (4)

(57) [Claims] An image processing apparatus that quantizes an image signal, propagates the quantization error to peripheral pixels, performs error diffusion, and inputs the quantized signal to a recording / display element to record or display. Adding means for adding an input image signal and an error signal dependent on the quantization error; and outputting an output signal of the adding means in accordance with density unevenness data representing density unevenness of recording or display by the recording / display element. , And quantizes the signal after the correction to record and display the signal.
Estimating the unevenness correction quantization means for supplying to the device, the density of the recorded or displayed by the recording / display device based signal quantized by the non-uniformity correction quantization means and the density unevenness data, corresponding to the concentration and concentration estimation means for outputting the density signal, an error calculation means for calculating an error between the output signal of the density signal and said adding means output from the density estimating means, based on the error calculated by said error calculation means the image processing apparatus characterized by comprising a means for generating said error signal Te. 2. An image processing apparatus which quantizes an image signal, propagates the quantization error to peripheral pixels to perform error diffusion, and inputs the quantized signal to a recording / display element to record or display. Adding means for adding an input image signal and an error signal dependent on the quantization error; and an output signal of the adding means having an amplitude equal to or smaller than a quantization level.
Signal adding means for adding a pattern signal consisting of a periodic pattern or a random pattern, and an output signal of the adding means after the pattern signal is added by the signal adding means, for recording or display density by the recording / display element. Correction is made in accordance with density unevenness data representing unevenness, and the signal after this correction is quantized to perform the recording /
And unevenness correction quantization means supplying to the display device, the unevenness of the density of the recorded or displayed by the recording / display device is estimated based on the correction quantization means by quantized signal and the density unevenness data in the concentration and concentration estimation means for outputting the corresponding density signal, an error calculation means for calculating an error between the output signal of the density signal and said adding means output from the density estimating means, the error calculated by said error calculation means Means for generating the error signal based on the error signal. 3. The image processing apparatus according to claim 1, wherein said unevenness correction quantization means, said density estimation means, and said error calculation means are configured as a table in one ROM. . 4. An image recorded or displayed when a predetermined signal is given to the recording / display element, and the density unevenness of the image is used as the density unevenness data. The image processing apparatus according to any one of the preceding claims.