JPH0120835B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to facsimile machines, copying machines, and other image processing devices that sample images and convert them into binary signals to create recorded images, that is, copies, and particularly relates to random noise removal and deciphering. The present invention relates to an image processing device that is capable of recording high-quality images due to improved performance.

Sample the image and convert it to a binary signal,
In facsimile etc. which makes a copy again,
The boundaries between white and black areas become vague, resulting in poor legibility, and random noise degrades image quality. Particularly, in the case of originals with low background density, such as newspapers, the occurrence of random noise is significant.

In addition, if legibility deteriorates or random noise occurs in this way, when encoding image signals for transmission, the code length becomes longer and the transmission time becomes longer. .

Conventionally, in order to improve readability, the binary data of a target pixel was corrected based on the state of surrounding pixels to increase the apparent MTF of the image.
A so-called MTF correction method (transfer characteristic correction method) has been proposed.

However, this MTF correction method does not have the effect of suppressing random noise, but rather tends to generate it. If random noise occurs, the legibility of the copy will deteriorate accordingly.

In addition, in an image processing device that uses an image signal converted to binary data, as a measure to remove random noise, it is possible to determine whether the target pixel is noise based on the similarity between the target pixel and pixels located around it. "Image noise processing method" to determine whether
55-27787) and "Image Processing Noise Removal Circuit" (Japanese Unexamined Patent Publication No. 54-152806) that determines whether a target pixel is black or white based on the information of the pixels above, below, left, or right of the target pixel.
(see Publication No. 2003), etc. are known.

Although these measures are effective for removing random noise, they do not improve legibility because they do not aim to emphasize the boundary between white and black areas.

Therefore, the image processing device of the present invention solves these difficulties in conventional correction methods, enhances the boundary between white and black areas to improve legibility, and removes random noise to achieve high quality. The purpose is to allow images to be copied.

To this end, in the image processing apparatus of the present invention, MTF correction is performed as preprocessing, and isolated point correction for removing random noise is performed as postprocessing.

As described above, the image processing apparatus of the present invention performs the known MTF correction as pre-processing, so in order to facilitate understanding, the MTF correction method will be explained first.

FIG. 1 is a conceptual processing system diagram for explaining the principle of a conventionally used MTF correction method. In the drawings, 1 is a line buffer matrix, 2 is an arithmetic circuit, and A to I are A/I.
In the D (analog/digital) conversion data, E is the target pixel, A to D and F to I are surrounding pixels, E^ is the corrected binary output signal of the target pixel E, and L _o-1 and
L _o and L _o+1 indicate line numbers of line buffers.

The surrounding pixels of the target pixel E are 8, A to D and F to I.
However, in practice, it is not necessary to modify the target pixel with all surrounding pixels. For example, it is sufficient to correct the four surrounding pixels B, D, F, and H on the top, bottom, left, and right, and depending on the correction data, the target pixel E is either white (=“0”) or black (=
“1”).

The correction function is αE=-β(B+D+F+H)...(1), and the corrected binary output signal E^ of the target pixel E
is calculated as E^=f(x) =f{αE−β(B+D+F+H)} (2). As an example, α=3, −β=−0.5
It is given as follows.

FIG. 2 is a block diagram showing an example of an MTF correction circuit. In the drawing, 3 is an amplifier, 4 is an A/D converter, 5 is a control logic circuit, 6 is a first latch circuit, 7 is a RAM, 8 is an address counter,
9 is a second latch circuit, 10 is an adder, and 11 is a
ROM, 12 indicates the third ripple circuit, and V _S
indicates the reference voltage.

For example, the image signal read by an image scanner is
It is input to amplifier 3 and converted to 4 by A/D converter 4.
Converted to bit binary data. In this case, the surrounding pixels B, D, F, and H are input in this order and written into the RAM 7 and further into the second latch circuit 9. Note that in the A/D converted binary data, (0, 0, 0, 0) is pure white, (1, 1,
1, 1) is considered pure black.

Therefore, the last binary data H (=H ₃ , H ₂ ,
When data (H ₁ , H ₀ ) are written, the data B, D, F of the previous line have already been latched in the second latch circuit 9 by the RAM 7.

Then, when the writing of the last data H to the second latch circuit 9 is completed, the adder 10
The calculation (B+D+F+H) in equation (2) above is executed and converted into 6-bit data.

ROM11 outputs A9 to A4 of this adder 10.
and the binary data A3 to A0 of the target pixel E, and output from the third latch circuit 12 to the outside as modified data E' of the target pixel E according to equation (2).

If no correction is performed, the boundary between white and black areas will be blurred and difficult to distinguish. Therefore, the copy has poor legibility, especially when it comes to white characters.

On the other hand, if processing is performed using the MTF correction circuit shown in Figure 2 above, the apparent MTF of the image will be
readability is improved.

However, in the case of originals with low background density, such as newspapers, random noise occurs, so it is not always possible to obtain high-quality copies.

Therefore, in the image processing device of the present invention, as already explained, MTF correction is performed as preprocessing.
Using the corrected binary data, isolated points are further corrected as post-processing to completely remove random noise.

As a result, a high-quality copy with good legibility and no random noise can be obtained.

The following FIG. 3 is a main part block diagram for explaining an embodiment of the image processing apparatus of the present invention. The reference numerals in the drawing are the same as those in FIG. 1, and 13 indicates a fourth latch circuit, 14 indicates an OR gate circuit, and 15 indicates an AND gate circuit.

In FIG. 3, the MTF correction circuit shown in FIG. 2 is used as pre-processing to obtain corrected binary data E^ according to equation (2), and as post-processing, surrounding pixels B, D, Isolated points are determined using arbitrary bits other than the LSB bits of F and H, here the MSB bit.

The A/D converted binary data of each pixel A to I is 4 bits, for example, B=B ₃ ,
The signals are inputted as B ₂ , B ₁ and B ₀ to the fourth latch circuit 13 in FIG.

In addition, the MTF-corrected binary data E^ is also
The signal is applied to this fourth latch circuit 13 from the MTF correction circuit shown in the figure.

The OR gate circuit 14 and the AND gate circuit 15 in FIG. 3 perform the calculation E=E^·(B ₃ +D ₃ +F ₃ +H ₃ )...(3).

Therefore, B ₃ , D ₃ , F ₃ , and H ₃ are all “0”, that is, B, D, F, and H are all (0,
1, 1, 1) In the following cases, isolated point corrected data E〓 is always E〓=0, regardless of E^.
In other words, it becomes white.

Therefore, compared to the case where only conventional MTF correction is performed, random noise is removed, the boundary between black and white is emphasized, and a high-quality image with good legibility can be obtained.

Also, by changing the OR logic in FIG. 3, that is, the OR gate circuit 14, B, D, F, H
It goes without saying that can be set to any value other than (0, 1, 1, 1).

In the above embodiment, the A/D converted binary data of each pixel is 4 bits. However, the same applies even in the case of 5 bits or more.

Similarly, the surrounding pixels are not only four, but also A~
It is also possible to use eight of D, F to I.

In addition, in the explanation of Fig. 3 above, the correction of 0 to K _W ,
That is, a correction has been described in which the color is forcibly rendered white when the value is less than a predetermined value. However, it goes without saying that correction for K _B <127, that is, correction for forcing the image to be completely black, is also possible when it is greater than or equal to a predetermined value.

As described above in detail, the image processing apparatus of the present invention performs known MTF correction as pre-processing, and performs isolated point correction for removing random noise as post-processing.

Therefore, according to the image processing apparatus of the present invention, it is possible to obtain a copy that is free from random noise and has good legibility because the boundary between white and black areas is emphasized. Furthermore, MTF correction emphasizes the boundaries and removing random noise strengthens the correlation of information.
Since unnecessary information is completely removed, the efficiency of encoding is improved, increasing the encoding speed and reducing the transmission time. Furthermore, the reading and recording devices can be similar to conventional ones, and the circuits for MTF correction and isolated point correction are LSI.
In addition to being suitable for use in various devices, the present invention can be used in common with various devices, resulting in many excellent effects such as being able to supply image processing devices at low cost.

[Brief explanation of drawings]

Figure 1 is a conceptual processing system diagram for explaining the principle of the conventionally used MTF correction method;
The figure is a block diagram showing an example of an MTF correction circuit, and FIG. 3 is a main part block diagram for explaining an embodiment of the image processing apparatus of the present invention. In the drawing, 1 is a line buffer matrix, 2 is an arithmetic circuit, 3 is an amplifier, 4 is an A/D converter, 5 is a control logic circuit, 6 is a first latch circuit, 7 is a RAM, 8 is an address counter, 9 is a second latch circuit, 10 is an adder, 11 is a ROM,
12 is a third latch circuit, and 13 is a fourth latch circuit.

Claims (1)

[Claims] 1. A means for sampling an image and generating a multiple-bit binary signal for one pixel by analog/digital conversion, and correction for the multiple-bit binary signal for one pixel generated by the means. An image processing device comprising an image correction means for correcting a target pixel according to the state of surrounding pixels and outputting a correction result as a binary signal,
The image correction means performs MTF correction processing.
MTF correction means and LSB of binary signal of surrounding pixels
An isolated point correction means is provided for correcting isolated points of the target pixel by inputting arbitrary bits other than bits, and a binary signal of a plurality of bits generated for one pixel is input to the MTF correction means. By performing MTF correction as pre-processing, and inputting the multi-bit binarized signal subjected to MTF correction processing by the MTF correction means to the isolated point correction means, as post-processing, the LSB of surrounding pixels is calculated. An image processing apparatus characterized in that isolated points are corrected by a combination of arbitrary bits other than bits.