JPH0214671A - Picture processing method - 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 [Technical Field] The present invention relates to an image processing method that can be applied to image processing apparatuses that use digital filters, such as digital copying machines and image processors.

[Prior art]

The matrix size of a digital filter used in a conventional image processing device is a normal 3×3 size.

By the way, a good color image means that it can faithfully reproduce the original, but since digital color copying machines handle digital signals sampled by a scanner, sampling errors that occur during image sampling,
There is MTF (transfer characteristic) deterioration due to quantization error, and MTF deterioration has already occurred in the sampled data.

This largely depends on the structure of the scanner.

That is, this is because the sampling mechanism is different in the main scanning direction and the sub-scanning direction. For example, in the main scanning direction, sensitivity unevenness of CC and D arrays, segment size of CCD array (sampling aperture size)
In addition, in the sub-scanning direction, jitter caused by vibration of the scanner causes deterioration of the MTI and data variations.

In this way, actual numbering data is not the same in the main and sub-scanning directions, but is sampled differently. However, as described above, the matrix size of the digital filter that performs MTF correction is 3×3 symmetrical, and proper filter processing is not performed.

〔the purpose〕

SUMMARY OF THE INVENTION An object of the present invention is to provide an image processing method that solves the problems of the above-mentioned conventional techniques, performs appropriate filter processing, and obtains images with good reproducibility.

〔composition〕

In order to achieve the above object, the present invention makes the size of the filter applied by the image processing section to the image data in the main and sub-scanning directions of the scanner different in the main-scanning direction and the sub-scanning direction, and the size of the filter The feature is that image processing is performed with the size set equal to or larger than the matrix size in the halftone processing section.

Hereinafter, one embodiment of the present invention will be specifically described based on the drawings.

Fig. 1 is a block diagram of the image processing section, Fig. 2 is a cross-sectional view of the image scanner, Fig. 3 is a block diagram of the document reading processing section, Fig. 4 is an MTF distribution diagram, Fig. 5 (a), b
l is an explanatory diagram of the smoothing filter, FIG. 6(a), (bl
is an explanatory diagram of the edge emphasis filter, Fig. 7 is an explanatory diagram of the digital filter, Fig. 8 is a block diagram of the line buffer and data latch of the digital filter, and Fig. 9 is a block diagram of the arithmetic circuit of the digital filter. be.

In FIG. 2, the image scanner 1 includes a contact glass 10 on which a document is placed, a light source 11, and a first scanning mirror 1.
2, a second scanning mirror 13, an imaging lens 14, a photoelectric conversion unit 15 that electrically reads reflected light from a document, and a light source 11
It is comprised of a lighting circuit 16 that drives the , a DC motor 17 that performs mechanical scanning (sub-scanning), a video processing circuit 2, and the like.

The Inogis 1-Yana 1 uses a fixed original scanning system to accommodate the diversity of originals, so that the optical path length of the original reflected light 19 is always constant during sub-scanning.
A first carriage Ca on which the light source 11 and the first scanning mirror 12 are mounted, and a second carriage Ca on which the second scanning mirror 13 is mounted are driven in the sub-scan by the DC motor 17 at a speed ratio of 2:1. Ru.

As in the document reading processing section shown in Figure 3, three C
The electrical signal (image density information) from the light-receiving surface of the CD15 is
The image is transferred to an image processing unit 30, which will be described later, and various processes are performed thereon. When the main scanning of one sub-scanning line by the CGD 15 is completed, the first carriage CaI and the second carriage Ca2 are driven for sub-scanning by the DC motor 17, and reading of the next sub-scanning line is started. In Figure 2, when the rask scan is completed to the right edge of the document, the DC motor 1
7, 1st gear position Cal % 2nd gear position C
a2 is return driven.

The rotational angular velocity of the DC motor 17 is fed back by the rotary encoder 13n and the speed signal generation circuit 23, and the sub-scanning speed control circuit 24 performs constant speed control and speed control such as starting and stopping.

The document reading process of the image scanner 1 described above is sequence-controlled by the scanner control circuit 25.

In FIG. 3, an A/D conversion circuit 21 converts an analog signal (image density information) read by the CCD 15 into an 8-bit digital signal. The read signal of the CCD 15 includes "distortion" caused by optical illumination unevenness or sensitivity variations of each element of the CCD 15. Therefore, in the shading correction circuit 22, the data after A/D conversion is corrected to remove the "distortion".

The image processing section 30 has the configuration shown in FIG. 1, and receives red, green, and blue (hereinafter abbreviated as R, G, and B) decomposition signals from the image scanner 1, and processes the printer's memory (not shown). 1D beam 0) ON10I='F
Output a signal.

In FIG. 1, a color conversion circuit 31 is a circuit that creates yellow, magenta, and cyan (hereinafter abbreviated as YM and C) from R, G, and B signals.
spectral distribution (optical filter for color separation used in the scanner) and Y, M, C (color toner used in the printer)
The γ correction circuit 32 changes the slope of the signal level to change the amount of toner adhering to the signal. Perform increases and decreases. In addition, the masking circuit 33 determines the ideal Y, M, and C spectral distribution, and the Y, M, and C colors of the toner to be used.
In order to correct the difference between the spectral reflection density of M and C,
This is a circuit that adjusts the level amount between Y, M, and C.
The UCR processing circuit 34 is a circuit that extracts equal amounts of each of the Y, M, and C signals as a Bk signal. The signal from the UCR processing circuit 34 is processed by a smoothing filter processing circuit 35, which is a halftone processing section, an edge emphasis filter processing circuit 36,
It is output to the density pattern circuit 37 and the multi-value processing circuit 38.

The T correction circuit 32 through the multi-value processing circuit 38 are controlled by a main body control circuit 39 and a synchronization control circuit 40.

The arithmetic processing in the smoothing filter processing circuit 35 and the edge enhancement filter processing circuit 36 is convolution integration, where the input image signal is S (x, y) and the filter is f.
(x, y), for example, a 3x3 matrix, S' (x+y) = S' 5S (x, y) books f
(x-x', y-y')dx'dy. However, * in the formula represents a convolution integral.

By the way, the results of measuring the image data are shown in Figure 4, and the deterioration of MTF in main and sub-scanning (main MT
F, the secondary MTF) is larger in the secondary MTF. In sub-scanning, since the scanner is constantly moving, the MTF of sampling data during movement is likely to deteriorate. In addition, there is local positional shift due to vibration, further deteriorating.

On the other hand, main scanning is sequential readout using a CCD array, and the signals are not degraded because they are moved electrically. However, as it approaches the sampling limit frequency determined by the numbering aperture of the CCD, it exhibits an MTF characteristic that gradually attenuates.

If this is all there is, the limit frequency is sufficiently high so there is no problem, but CCDs have uneven sensitivity, which becomes noise and is added to the signal. This appears regularly in the main scanning direction and is a major factor in deteriorating image quality.

That is, it can be considered that data variation and roughness occur in the main scanning direction, and data blur occurs in the sub-scanning direction.

As mentioned above, since the MTF is different in the main and sub-scanning directions, in order to obtain a homogeneous image, it is necessary to correct the MTF in the main and sub-scanning directions so that the MTF is equal to that between the lines. for that purpose,
It is preferable that the matrix size of the filter is different in the main and sub-scanning directions, for example, 3 (sub-scanning direction)×5 (main-scanning direction), similarly 3×7, 5×7, etc.

Further, the filter size has no meaning unless it is equal to or larger than the matrix size used in the halftone processing units 35 to 38. This is because in halftone processing, the shading of an image in a matrix is determined by the arrangement of darkness values in the matrix, so the MTF effect of the previous filter processing disappears.

Next, specific filter processing will be explained. This digital filter has a size of 3 (sub-scanning direction) x 5 (main-scanning direction). It is assumed that a filter 50 in which coefficients are arranged in grids as shown in FIG. 7 assigns numbers ``■'' to ``■'' to each image data.

The center ■ is the pixel of interest, and image data for this position is calculated. As shown in FIG. 8, line buffers LB 1° LB2 and 15 (
= 3×5) A data latch l-1” LIS for pixels is provided, and as shown in FIG.
The image data of the pixel is fa in FIGS. 5 and 6) or (
The coefficients shown in b) are input to an arithmetic circuit that performs a product-sum operation with the smoothing filter 35 or the edge emphasis filter 36.

In this circuit, pixels having the same distance from the pixel of interest (2) are grouped and sequentially added. Pixel ■, ■, ■.

■, ■, @l are adders 51, 52, 53, 58, 511
and the result is multiplied by the numerical value input in the multiplier 61.

For example, the filter If so, the value of @ is 1 and multiplied by 1.

Pixel ■, ■, 0.0. ■, o, ■, ■ are adders 54 to 5
7, 59, 510, 512, and the result is multiplier 6.
In step 2, the value entered in ■ is multiplied by 2.

Then, the sum of 15 pixels including the value inputted to ◎ in the multiplier 63, that is, the pixel 8 multiplied by 4, is calculated by the adder 51.
4 and multiplied by 1/26 in a multiplier 64. In this way, the final calculation result is obtained. This output becomes the output value of one pixel, and is sequentially applied to all pixels to perform edge enhancement or smoothing processing.

The optimum value of the matrix size of the filter is changed depending on the sampling density.

In this way, optimal MTF correction is performed in each of the main and sub-scanning directions, and isotropic and uniform image quality is obtained.

〔effect〕

As explained above, according to the image processing method according to the present invention, it is possible to perform good filter processing and obtain an image with good reproducibility.

[Brief explanation of the drawing]

The drawings show one embodiment of the present invention, and FIG. 1 is a block diagram of an image processing section, FIG. 2 is a sectional view of an image scanner, and FIG.
The figure is a block diagram of the document reading processing section, and the first figure is the MTF.
distribution diagram, Figure 5 (al, (bl is smoothing) / Ilter explanatory diagram, Figure 6 + a1. (bl is an explanatory diagram of edge emphasis filter, Figure 7 is an explanatory diagram of digital filter, Figure 8
The figure is a block diagram of the line buffer and data latch of the digital filter, and FIG. 9 is a block diagram of the arithmetic circuit of the digital filter. 1... Scanner, 30... Image processing section, 35, 36
゜37.38... Halftone processing section, 35.36... Filter. 40 □ Shusaka Aki 5th prisoner (al (b) Figure 6 (Q) 121 (b) -11112-1± Figure 7 Figure 9

Claims (1)

[Claims] The size of the filter applied by the image processing unit to the image data in the main and sub-scanning directions of the scanner is different in the main-scanning direction and the sub-scanning direction, and the size of the filter is the same as the matrix size in the halftone processing unit. An image processing method characterized by performing image processing with the above settings.